October 2008 - Olena-patches

r2674: Enhance the build-system
by Maxime van Noppen 24 Oct '08

24 Oct '08

From: Maxime van Noppen <yabo(a)lrde.epita.fr> To: olena-patches(a)lrde.epita.fr Subject: r2674: Enhance the build-system URL: https://svn.lrde.epita.fr/svn/oln/branches/cleanup-2008/milena/sandbox ChangeLog: 2008-10-24 Maxime van Noppen <yabo(a)lrde.epita.fr> Enhance the build-system. * Makefile: Make a nice and usefull Makefile. --- Makefile | 48 +++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 43 insertions(+), 5 deletions(-) Index: branches/cleanup-2008/milena/sandbox/classif/Makefile =================================================================== --- branches/cleanup-2008/milena/sandbox/classif/Makefile (revision 2673) +++ branches/cleanup-2008/milena/sandbox/classif/Makefile (revision 2674) @@ -1,7 +1,45 @@ -CCFLAGS=-I../.. -O1 -DNDEBUG +INCLUDES=-I../.. -all: - g++ iccvg04.cc $(CCFLAGS) +SRC=iccvg04.cc +TARGET=iccvg -debug: - g++ iccvg04.cc -I../../ -O0 -g3 -ggdb3 +TARGET_DBG=iccvg_dbg + +IMG?=../../img/lena.ppm +DIV?=8 +LAMBDA?=10 + +LOG=> stdout.log 2> stderr.log + +all: $(TARGET) + +$(TARGET): $(SRC) + g++ $(INCLUDES) -O1 -DNDEBUG $(SRC) -o $(TARGET) + +debug: $(TARGET_DBG) + +iccvg_dbg: $(SRC) + g++ $(INCLUDES) -O0 -g3 -ggdb3 $(SRC) -o $(TARGET_DBG) + +.PHONY:clean check check-debug valgrind + +clean: + rm -f $(TARGET) + rm -f $(TARGET_DBG) + +check-debug: $(TARGET_DBG) + ./iccvg_dbg $(IMG) $(DIV) $(LAMBDA) $(LOG) + cat stdout.log + display out.ppm & + +check: $(TARGET) + ./iccvg $(IMG) $(DIV) $(LAMBDA) $(LOG) + cat stdout.log + display out.ppm & + +valgrind: + valgrind --log-file=valgrind.log ./iccvg_dbg ../../img/lena.ppm $(DIV) $(LAMBDA) $(LOG) + +gdb: $(TARGET_DBG) + echo "run ../../img/lena.ppm $(DIV) $(LAMBDA)" > gdb.cmd + gdb $(TARGET_DBG) -x gdb.cmd

1 0

cleanup-2008 2673: Add bijective functors, 2-way functors and meta-fun.
by Alexandre Abraham 24 Oct '08

24 Oct '08

https://svn.lrde.epita.fr/svn/oln/branches/cleanup-2008/milena/sandbox Index: ChangeLog from Alexandre Abraham <abraham(a)lrde.epita.fr> Add bijective functors, 2-way functors and meta-fun. * abraham/tests/morpho/test_component_tree.cc: Remove This test was outdated. * abraham/tests/morpho/test_watershed.cc: Clean. * abraham/tests/morpho/test_watershed_topo.cc: Clean. * abraham/tests/fun/meta: New. * abraham/tests/fun/meta/red.cc: New (mln::meta::red) : meta function to access red in rgb. * abraham/tests/core/image/thru_norm.cc: New test for norm functor. * abraham/tests/core/image/thru_v2v.cc: New test for abs functor. * abraham/tests/core/image/thru.cc: Remove. * abraham/tests/core/image/thru_const.cc: New test of writing in a const image. * abraham/tests/core/image/thru_v2w2v.cc: New test for cosinus functor. * abraham/tests/core/concept: New. * abraham/tests/core/concept/test.cc: New. * abraham/mln/morpho/najman_wst.hh: New M-W watershed. * abraham/mln/morpho/basic_najman.hh: Obsolete, will eventually be removed. * abraham/mln/morpho/topo_wst.hh: New 1 pass topological watershed. * abraham/mln/core/image/thru.hh: (mln::thru): see an image throught a function. * abraham/mln/core/image/shell.hh: New (mln::shell): value shell. * abraham/mln/core/concept/function.hh: Add bijective and 2-way function. * abraham/mln/core/concept/meta_fun.hh: New Add concept for meta function. * abraham/mln/fun/meta: New. * abraham/mln/fun/meta/red.hh: New red fun object. * abraham/mln/fun/v2w_w2v/norm.hh: 2-way functor. * vigouroux/tests.cc: Rename image_if_value > image_if. * vigouroux/color/rgb_to_cmy.hh: Rename image_if_value > image_if. * vigouroux/color/rgb_to_xyz.hh: Rename image_if_value > image_if. * vigouroux/color/rgb_to_hsi.hh: Rename image_if_value > image_if. * vigouroux/color/rgb_to_yuv.hh: Rename image_if_value > image_if. abraham/mln/core/concept/function.hh | 2 abraham/mln/core/concept/meta_fun.hh | 83 ++ abraham/mln/core/image/shell.hh | 118 ++++ abraham/mln/core/image/thru.hh | 40 - abraham/mln/fun/meta/red.hh | 45 + abraham/mln/fun/v2w_w2v/norm.hh | 4 abraham/mln/morpho/basic_najman.hh | 196 +++--- abraham/mln/morpho/najman_wst.hh | 791 ++++++++++++++++++++++++++++ abraham/mln/morpho/topo_wst.hh | 744 ++++++++++++++++++++++++++ abraham/tests/core/concept/test.cc | 75 ++ abraham/tests/core/image/thru.cc | 56 - abraham/tests/core/image/thru_const.cc | 64 ++ abraham/tests/core/image/thru_norm.cc | 58 ++ abraham/tests/core/image/thru_v2v.cc | 56 + abraham/tests/core/image/thru_v2w2v.cc | 64 ++ abraham/tests/fun/meta/red.cc | 60 ++ abraham/tests/morpho/test_component_tree.cc | 230 -------- abraham/tests/morpho/test_watershed.cc | 60 -- abraham/tests/morpho/test_watershed_topo.cc | 8 vigouroux/color/rgb_to_cmy.hh | 8 vigouroux/color/rgb_to_hsi.hh | 2 vigouroux/color/rgb_to_xyz.hh | 14 vigouroux/color/rgb_to_yuv.hh | 8 vigouroux/tests.cc | 6 24 files changed, 2311 insertions(+), 481 deletions(-) Index: abraham/tests/morpho/test_watershed.cc --- abraham/tests/morpho/test_watershed.cc (revision 2672) +++ abraham/tests/morpho/test_watershed.cc (working copy) @@ -7,7 +7,7 @@ #include <mln/io/pgm/load.hh> #include <mln/io/pgm/save.hh> -#include <mln/morpho/basic_najman.hh> +#include <mln/morpho/najman_wst.hh> #include <string> #include <iostream> @@ -22,15 +22,15 @@ // #define TEST - io::pgm::load(input, "./images/test_watershed.pgm"); + // io::pgm::load(input, "./images/test_watershed.pgm"); // io::pgm::load(input, "./images/little_test.pgm"); - // io::pgm::load(input, "./images/test_2.pgm"); + io::pgm::load(input, "./images/test.pgm"); // io::pgm::load(input, "./images/lena_light.pgm"); // io::pgm::load(mverif, "./images/result_m_watershed.pgm"); // io::pgm::load(wverif, "./images/result_topo_watershed.pgm"); - morpho::basic_najman< image2d<int_u8>, neighb2d> n(input, c4()); + morpho::najman_wst< image2d<int_u8>, neighb2d> n(input, c4()); /* image2dint::fwd_piter it(input.domain()); @@ -43,60 +43,8 @@ // io::tikz::save(input, "start.tex"); - std::cout << "Building Component tree..." << std::endl; - n.go(); - std::cout << "M-Watershed" << std::endl; - - n.m_watershed(); - - io::pgm::save(n.pima, "int.pgm"); - - // io::tikz::save(n.pima, "step.tex"); - -#ifdef TEST - - bool equal = true; - - equal = n.pima == mverif; - - if (!equal) - { - std::cout << "M-Watershed broken" << std::endl; - return 1; - } - else - std::cout << "M-Watershed good as chocolate ice cream" << std::endl; - -#endif - - std::cout << "W-watershed" << std::endl; - - n.w_watershed(); - // io::pgm::save(n.pima, "int.pgm"); - // std::cout << "~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~" << std::endl; - // n.w_watershed(); - - // for_all(it) - // n.pima(it) = n.pima(it) * 17; - -#ifdef TEST - equal = n.pima == wverif; - - if (!equal) - { - std::cout << "W-Watershed broken" << std::endl; - return 1; - } - else - std::cout << "W-Watershed good as chocolate ice cream" << std::endl; -#endif - - // io::tikz::save(n.pima, "end.tex"); - - // n.image_output(n.pima); - io::pgm::save(n.pima, "out.pgm"); return 0; Index: abraham/tests/morpho/test_watershed_topo.cc --- abraham/tests/morpho/test_watershed_topo.cc (revision 2672) +++ abraham/tests/morpho/test_watershed_topo.cc (working copy) @@ -9,7 +9,7 @@ #include <mln/io/pgm/load.hh> #include <mln/io/pgm/save.hh> -#include <mln/morpho/basic_najman.hh> +#include <mln/morpho/topo_wst.hh> #include <string> #include <iostream> @@ -32,7 +32,7 @@ // io::pgm::load(input, "./images/test_watershed.pgm"); // io::pgm::load(input, "./images/little_test.pgm"); - io::pgm::load(input, "./images/test_4.pgm"); + io::pgm::load(input, "./images/test.pgm"); // io::pgm::load(input, "../../img/dots.pgm"); //io::pgm::load(input, "./images/+irm6.pgm"); @@ -40,7 +40,7 @@ // io::pgm::load(mverif, "./images/result_m_watershed.pgm"); // io::pgm::load(wverif, "./images/result_topo_watershed.pgm"); - morpho::basic_najman< image2d<int_u8>, neighb2d> n(input, c4()); + morpho::topo_wst< image2d<int_u8>, neighb2d> n(input, c4()); /* image2dint::fwd_piter it(input.domain()); @@ -53,7 +53,7 @@ // io::tikz::save(input, "start.tex"); - n.gotopo(); + n.go(); io::pgm::save(n.pima, "out.pgm"); Index: abraham/tests/fun/meta/red.cc --- abraham/tests/fun/meta/red.cc (revision 0) +++ abraham/tests/fun/meta/red.cc (revision 0) @@ -0,0 +1,60 @@ +#include <mln/fun/meta/red.hh> +#include <mln/core/image/image2d.hh> +#include <mln/core/image/thru.hh> + +namespace mln +{ + + template <class T> + struct rgb + { + T red() const { return r; } + T& red() { return r; } + T r; + }; + + template <class T> + struct function< meta::red, rgb<T> > : public Function_v2w_w2v<function< meta::red, rgb<T> > > + { + typedef T result; + T read(const rgb<T>& c) + { + std::cout << "read red rgb<T>" << std::endl; + return c.red(); + } + + typedef T& lresult; + T& write(rgb<T>& c) + { + std::cout << "write red rgb<T>" << std::endl; + return c.red(); + } + }; +} + +int main () +{ + typedef mln::rgb<int> C; + mln::image2d<C> i(3, 2, 0); + C c; + c.r = 51; + i(mln::point2d(0,0)) = c; + c.r = 23; + i(mln::point2d(0,1)) = c; + c.r = 43; + i(mln::point2d(1,0)) = c; + c.r = 0; + i(mln::point2d(1,1)) = c; + c.r = 65; + i(mln::point2d(2,0)) = c; + c.r = 1; + i(mln::point2d(2,1)) = c; + + mln::thru<mln::meta::red, mln::image2d<C> > out(i); + + for_all (p) + std::cout << out(p) << std::endl; + + +} + Index: abraham/tests/core/image/thru_norm.cc --- abraham/tests/core/image/thru_norm.cc (revision 0) +++ abraham/tests/core/image/thru_norm.cc (revision 0) @@ -0,0 +1,58 @@ +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +# include <mln/core/image/image2d.hh> +# include <mln/core/image/thru.hh> +# include <mln/fun/v2w_w2v/norm.hh> +# include <mln/algebra/vec.hh> +# include <mln/level/fill.hh> +# include <mln/core/image/violent_cast_image.hh> + + +int main () +{ + + using namespace mln; + typedef image2d<algebra::vec<2, double> > I; + + I ima(3, 2, 0); + + ima(point2d(0,0)).set (1, 1); + ima(point2d(0,1)).set (1, 3); + ima(point2d(1,0)).set (4, 4); + ima(point2d(1,1)).set (-1, 3); + ima(point2d(2,0)).set (23, 23); + ima(point2d(2,1)).set (3, 1); + + thru<mln::fun::v2w_w2v::l1_norm<algebra::vec<2, double>, double>, I > out(ima); + level::fill(out, 1); + + box_fwd_piter_<point2d> p(ima.domain()); + + for_all (p) + std::cout << ima(p) << std::endl; +} Index: abraham/tests/core/image/thru_v2v.cc --- abraham/tests/core/image/thru_v2v.cc (revision 0) +++ abraham/tests/core/image/thru_v2v.cc (revision 0) @@ -0,0 +1,56 @@ +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +# include <mln/core/image/image2d.hh> +# include <mln/core/image/thru.hh> +# include <mln/fun/v2v/abs.hh> + +int main () +{ + + using namespace mln; + + typedef image2d<int> I; + + int vs[6][5] = { + + { -3, -3, -4, -4, -4 }, + { -2, -1, -1, -1, -1 }, + { -1, -4, -4, -4, -1 }, + { -1, -4, -3, -4, -1 }, + { -1, -4, -5, -3, -1 }, + { -1, -1, -1, -1, -1 } + + }; + + image2d<int> ima(make::image2d(vs)); + thru<mln::fun::v2v::abs<int>, image2d<int> > out(ima); + + box_fwd_piter_<point2d> p(ima.domain()); + for_all (p) + mln_assertion (out(p) >= 0); +} Index: abraham/tests/core/image/thru_const.cc --- abraham/tests/core/image/thru_const.cc (revision 0) +++ abraham/tests/core/image/thru_const.cc (revision 0) @@ -0,0 +1,64 @@ +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +# include <mln/core/image/image2d.hh> +# include <mln/core/image/thru.hh> +# include <mln/fun/v2w2v/cos.hh> + +int main () +{ + + using namespace mln; + + typedef const image2d<double> I; + + double vs[6][5] = { + + { 12, -3, 8, -4, 6 }, + { -2, 22, -1, 45, -1 }, + { -1, -4, -4, -4, -1 }, + { -1, -4, -3, -4, -1 }, + { -1, -4, -5, -3, -1 }, + { -1, -1, -1, -1, -1 } + + }; + + I ima(make::image2d(vs)); + thru<mln::fun::v2w2v::cos<double>, I > out(ima); + + double i = 0; + + box_fwd_piter_<point2d> p(ima.domain()); + for_all (p) + { + out(p) = i; + i += 1./40.; + } + + for_all (p) + std::cout << out(p) << std::endl; +} Index: abraham/tests/core/image/thru_v2w2v.cc --- abraham/tests/core/image/thru_v2w2v.cc (revision 0) +++ abraham/tests/core/image/thru_v2w2v.cc (revision 0) @@ -0,0 +1,64 @@ +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +# include <mln/core/image/image2d.hh> +# include <mln/core/image/thru.hh> +# include <mln/fun/v2w2v/cos.hh> + +int main () +{ + + using namespace mln; + + typedef image2d<double> I; + + double vs[6][5] = { + + { 12, -3, 8, -4, 6 }, + { -2, 22, -1, 45, -1 }, + { -1, -4, -4, -4, -1 }, + { -1, -4, -3, -4, -1 }, + { -1, -4, -5, -3, -1 }, + { -1, -1, -1, -1, -1 } + + }; + + image2d<double> ima(make::image2d(vs)); + thru<mln::fun::v2w2v::cos<double>, image2d<double> > out(ima); + + double i = 0; + + box_fwd_piter_<point2d> p(ima.domain()); + for_all (p) + { + out(p) = i; + i += 1./40.; + } + + for_all (p) + std::cout << out(p) << std::endl; +} Index: abraham/tests/core/concept/test.cc --- abraham/tests/core/concept/test.cc (revision 0) +++ abraham/tests/core/concept/test.cc (revision 0) @@ -0,0 +1,75 @@ +#include <iostream> +#include <mln/core/concept/meta_fun.hh> + +using namespace mln; + +meta::red red; // fun-obj + + + +template <class T> +struct rgb +{ + T red() const { return r; } + T& red() { return r; } + T r; +}; + + + +template <class T> +struct fun< meta::red, rgb<T> > : Function_v2v< fun< meta::red, rgb<T> > > +{ + typedef T result; + T read(const rgb<T>& c) + { + std::cout << "read red rgb<T>" << std::endl; + return c.red(); + } + + typedef T& lresult; + T& write(rgb<T>& c) + { + std::cout << "write red rgb<T>" << std::endl; + return c.red(); + } +}; + + +/* + + morpher apply(M m, I ima) + { + M et value(I) -> F -> nature + } + + */ + + + +int main() +{ + typedef rgb<int> C; + C c; + c.r = 51; + + int r = red(c); + std::cout << r << std::endl; + + + // norm(v) = 1; + + + // norm( ima(p) ) = n; + + // norm( v& ) = n; + // v = n * v / norm(v); + + // norm( v& ) = 1; + + + + // ima' = apply(norm, ima) + // ima'(p) = 1; + +} Index: abraham/mln/morpho/najman_wst.hh --- abraham/mln/morpho/najman_wst.hh (revision 0) +++ abraham/mln/morpho/najman_wst.hh (revision 0) @@ -0,0 +1,791 @@ +// Copyright (C) 2008 EPITA Research and Development Laboratory (LRDE) +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +#ifndef MLN_MORPHO_NAJMAN_WST_HH +# define MLN_MORPHO_NAJMAN_WST_HH + + + +#include <mln/level/sort_psites.hh> +#include <mln/level/fill.hh> +#include <mln/core/image/image2d.hh> +#include <mln/core/site_set/p_set.hh> +#include <mln/estim/min_max.hh> +#include <mln/math/sqr.hh> +#include <mln/util/greater_psite.hh> +#include <mln/util/ord.hh> + +#include <mln/core/site_set/p_priority.hh> +#include <mln/core/site_set/p_queue_fast.hh> + +#include <queue> +#include <vector> +#include <map> + +namespace mln +{ + namespace morpho + { + + template <class I, class N> + struct najman_wst + { + + // Typedefs + // -------- + typedef mln_site(I) site; + + // Component tree management + // ------------------------- + private: + struct node { + mln_value(I) level; + int area; + int highest; + typedef mln_site(I) site; + // Can't modify the sites in a p_array + // p_array<mln_site(I)> children; + std::vector<site> children; + + void addChildren(const node& n) + { + // typename p_array<mln_site(I)>::fwd_piter it(n.children); + // for (it.start(); + // it.is_valid(); + // it.next()) + // children.append(it.to_site()); + for (unsigned i=0; i < n.children.size(); ++i) + children.push_back(n.children[i]); + } + + void addChild(const site p) + { + // children.append(n); + children.push_back(p); + } + }; // struct node + + const mln_exact(N)& nbh; + mln_ch_value(I, site) Par_node; + mln_ch_value(I, site) Par_tree; + + mln_ch_value(I, int) Rnk_tree; + mln_ch_value(I, int) Rnk_node; + mln_ch_value(I, site) subtreeRoot; + mln_ch_value(I, node) nodes; + site Root; + + void MakeSet_tree(site x); + void MakeSet_node(site x); + site Find_tree(site x); + void swap(site& x, site& y); + site Find_node(site x); + void BuildComponentTree(); + site MergeNode(site& node1, site& node2); + site Link_tree(site x, site y); + site Link_node(site x, site y); + node MakeNode(int level); + + + // Watershed algorithm + // ------------------- + + private: + mln_ch_value(I, bool) isproc; + + // Ctor + public: + najman_wst(const Image& i, + const Neighborhood<N>& n); + + + // Result + public: + mln_exact(I) pima; + + public: + void go(); + + private: + site min (p_set<site>& components); + site max (p_set<site>& components); + bool highest_fork (p_set<site>& components); + bool highest_fork (p_set<site>& components, site &r); + bool m_destructible(site p); + bool w_destructible(site p); + bool m_destructible(site p, site &r); + bool w_destructible(site p, site &r); + void m_watershed (); + void w_watershed (); + + // Optimized LCA Algorithm + private: + int *euler; + int *depth; + int ctree_size; + std::map<site, int, mln::util::ord<site> > pos; + site *repr; + + int *minim; + int **Minim; + + void compute_ctree_size (site p); + void build_euler_tour_rec(site p, int &position, int d); + void build_euler_tour(); + void build_minim (); + site lca (site a, site b); + void removeOneSonNodes(site *p, mln_ch_value(I, site) &newPar_node); + void compressTree(); + }; // struct najman_wst + +# ifndef MLN_INCLUDE_ONLY + + // Ctor + template <class I, class N> + najman_wst<I, N>::najman_wst(const Image& i, + const Neighborhood<N>& n) + : pima(exact(i)), + nbh(exact(n)), + Par_node(exact(i).domain(), exact(i).border()), + Par_tree(exact(i).domain(), exact(i).border()), + Rnk_tree(exact(i).domain(), exact(i).border()), + Rnk_node(exact(i).domain(), exact(i).border()), + subtreeRoot(exact(i).domain(), exact(i).border()), + nodes(exact(i).domain(), exact(i).border()), + isproc(exact(i).domain(), exact(i).border()) + { + } + + template <class I, class N> + void najman_wst<I, N>::MakeSet_tree(site x) + { + Par_tree(x) = x; + Rnk_tree(x) = 0; + } + + template <class I, class N> + void najman_wst<I, N>::MakeSet_node(site x) + { + Par_node(x) = x; + Rnk_node(x) = 0; + } + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::Find_tree(site x) + { + if (Par_tree(x) != x) + Par_tree(x) = Find_tree(Par_tree(x)); + return Par_tree(x); + } + + template <class I, class N> + void najman_wst<I, N>::swap(site& x, site& y) + { + site memo = x; + x = y; + y = memo; + } + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::Find_node(site x) + { + if (Par_node(x) != x) + Par_node(x) = Find_node(Par_node(x)); + return Par_node(x); + } + + template <class I, class N> + void najman_wst<I, N>::go() + { + BuildComponentTree(); + compressTree(); + build_euler_tour(); + build_minim(); + m_watershed(); + w_watershed(); + } + + template <class I, class N> + void najman_wst<I, N>::BuildComponentTree() + { + // Init: + + // Sort the sites in increasing order + p_array<mln_site(I)> S; + S = level::sort_psites_increasing(pima); + + // Clear the marker map + level::fill(isproc, false); + for (int ip = 0; ip < int(S.nsites()); ++ip) + { + site p = S[ip]; + MakeSet_node(p); + MakeSet_tree(p); + // if (subtreeRoot.hold(p)) + subtreeRoot(p) = p; + // if (nodes.hold(p)) + nodes(p) = MakeNode(pima(p)); + } + + + + typename p_array<mln_site(I)>::fwd_piter ip(S); + for_all(ip) + { + site p = ip.to_site(); + + site curCanonicalElt = Find_tree(p); + site curNode = Find_node(subtreeRoot(curCanonicalElt)); + + mln_niter(N) q(nbh, ip); + for_all(q) + if (pima.has(q) and isproc(q) and pima(q) <= pima(p)) + { + site adjCanonicalElt = Find_tree(q); + site adjNode = Find_node(subtreeRoot(adjCanonicalElt)); + if (curNode != adjNode) + { + if (nodes(curNode).level == nodes(adjNode).level) + curNode = MergeNode(adjNode, curNode); + else + { + nodes(curNode).addChild(adjNode); + nodes(curNode).area += nodes(adjNode).area; + nodes(curNode).highest += nodes(adjNode).highest; + } + } + + curCanonicalElt = Link_tree(adjCanonicalElt, curCanonicalElt); + subtreeRoot(curCanonicalElt) = curNode; + } + isproc(p) = true; + } + // Pour garder une map de correspondance site <-> local_root + // for (int ip = 0; ip < int(S.size()); ++ip) + // { + // site p = S[ip]; + // M(p) = Find_node(p); + // } + + mln_piter(I) r(Par_node.domain()); + for_all(r) + Par_node(r) = Find_node(r); + + Root = subtreeRoot(Find_tree(Find_node(site(0, 0)))); + } + + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::MergeNode(site& node1, site& node2) + { + site tmpNode = Link_node(node1, node2); + site tmpNode2; + if (tmpNode == node2) + { + nodes(node2).addChildren(nodes(node1)); + tmpNode2 = node1; + } + else + { + nodes(node1).addChildren(nodes(node2)); + tmpNode2 = node2; + } + nodes(tmpNode).area += nodes(tmpNode2).area; + nodes(tmpNode).highest += nodes(tmpNode2).highest; + return tmpNode; + } + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::Link_tree(site x, site y) + { + if (Rnk_tree(x) > Rnk_tree(y)) + swap(x, y); + else + if (Rnk_tree(x) == Rnk_tree(y)) + Rnk_tree(y) += 1; + Par_tree(x) = y; + return y; + } + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::Link_node(site x, site y) + { + if (Rnk_node(x) > Rnk_node(y)) + swap(x, y); + else + if (Rnk_node(x) == Rnk_node(y)) + Rnk_node(y) += 1; + Par_node(x) = y; + return y; + } + + template <class I, class N> + typename najman_wst<I, N>::node najman_wst<I, N>::MakeNode(int level) + { + node n; + n.level = level; + n.area = 1; + n.highest = level; + return n; + } + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::min (p_set<site>& components) + { + if (components.nsites() == 0) + return site(-1, -1); + + typename p_set<site>::fwd_piter it(components); + site min = components[0]; + + for_all(it) + if (pima(it.to_site()) < pima(min)) + min = it.to_site(); + + return min; + } + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::max (p_set<site>& components) + { + if (components.nsites() == 0) + return site(-1, -1); + + typename p_set<site>::fwd_piter it(components); + site max = components[0]; + + for_all(it) + if (pima(it.to_site()) > pima(max)) + max = it.to_site(); + + return max; + } + + + template <class I, class N> + bool najman_wst<I, N>::highest_fork (p_set<site>& components, site &r) + { + if (components.nsites() == 0) + { + std::cerr << "highest fork : empty set" << std::endl; + return false; + } + + site + m = min(components), + hfork = m; + + typename p_set<site>::fwd_piter it(components); + for_all(it) + { + // Can't remove the site from the set + if (it.to_site() == m) + continue; + + site c = lca(hfork, it.to_site()); + if (c != it.to_site()) + hfork = c; + } + + if (nodes(m).level == nodes(hfork).level) + return false; + + r = hfork; + return true; + } + + template <class I, class N> + bool najman_wst<I, N>::w_destructible(site p, site &r) + { + mln_niter(N) q(nbh, p); + p_set<site> v; + + for_all(q) + if (pima.has(q) && pima(q) < pima(p)) + v.insert(Par_node(q)); + + if (v.nsites() == 0) + return false; + + site hf; + bool is_hf = highest_fork(v, hf); + + if (!is_hf) { + r = min(v); + return true; + } + + if (nodes(hf).level <= pima(p)) { + r = hf; + return true; + } + return false; + } + + template <class I, class N> + bool najman_wst<I, N>::highest_fork (p_set<site>& components) { + site i; + return highest_fork(components, i); + } + + template <class I, class N> + bool najman_wst<I, N>::m_destructible(site p) { + site i; + return m_destructible(p, i); + } + + template <class I, class N> + bool najman_wst<I, N>::w_destructible(site p) { + site i; + return w_destructible(p, i); + } + + template <class I, class N> + bool najman_wst<I, N>::m_destructible(site p, site &r) + { + mln_niter(N) q(nbh, p); + p_set<site> v = p_set<site>(); + + for_all(q) + if (pima.has(q) && pima(q) < pima(p)) + v.insert(Par_node(q)); + + if (v.nsites() == 0) + return false; + + // if (nodes(min(v)).children.nsites() != 0) + if (nodes(min(v)).children.size() != 0) + return false; + + bool is_hf = highest_fork(v); + + if (!is_hf) { + r = min(v); + return true; + } + + return false; + } + + + template <class I, class N> + void najman_wst<I, N>::m_watershed () + { + typedef + // p_priority_queue<site, mln_value(I) > + std::priority_queue<site, std::vector<site>, util::greater_psite > + ordered_queue_type; + + + ordered_queue_type l(util::make_greater_psite(pima)); + + // Clear the marker map + level::fill(isproc, false); + mln_piter(I) it(pima.domain()); + + for_all(it) + if (m_destructible(it.to_site())) + { + l.push(it.to_site()); + isproc(it.to_site()) = true; + } + + site p, i; + while (!l.empty()) + { + // Extract priority queue + p = l.top(); + l.pop(); + + // unmark p + isproc(p) = false; + + bool is_m = m_destructible(p, i); + + if (is_m) + { + pima(p) = nodes(i).level; + Par_node(p) = i; + + mln_niter(N) q(nbh, p); + for_all(q) + if (pima.has(q) && !isproc(q)) + if (m_destructible(q)) + { + // Add priority queue + l.push(q); + + // mark q + isproc(q) = true; + } + } + } + } + + + template <class I, class N> + void najman_wst<I, N>::w_watershed () + { + mln::p_priority< mln_value(I), p_queue_fast<site> > L; + + mln_value(I) max = level::compute(accu::meta::max(), pima); + + // I K(pima.domain(), pima.border()); + mln_ch_value(I, unsigned) K(pima.domain(), pima.border()); + mln_ch_value(I, site) H(pima.domain(), pima.border()); + + // For All p of E do + mln_piter(I) it(pima.domain()); + for_all(it) + { + site p = it.to_site(); + site i; + + // i <- W-Destructible(p) + bool is_w = w_destructible(p, i); + + // If i != infinity then + if (is_w) + { + L.push(max - nodes(i).level, p); + K(p) = nodes(i).level; + H(p) = i; + } + } + + while (!L.is_empty()) + { + mln_value(I) k = max - L.highest_priority(); + + site p = L.front(); + L.pop(); + + if (K(p) == k) + { + pima(p) = k; + Par_node(p) = H(p); + + mln_niter(N) q(nbh, p); + for_all(q) + if (pima.has(q) && k < pima(q)) + { + site i; + bool is_w = w_destructible(q, i); + if (!is_w) + K(q) = 10000; // FIXME : supposed to be infinity... + else + if (K(q) != nodes(i).level) + { + L.push(max - nodes(i).level, q); + K(q) = nodes(i).level; + H(q) = i; + } + } + } + } + } + + + template <class I, class N> + void najman_wst<I, N>::compute_ctree_size (site p) + { + ctree_size += 1; + node& n = nodes(p); + + // typename p_array<mln_site(I)>::fwd_piter it(n.children); + // for_all(it) + // compute_ctree_size(it.to_site()); + + for (unsigned i=0; i < n.children.size(); ++i) + compute_ctree_size(n.children[i]); + } + + + template <class I, class N> + void najman_wst<I, N>::build_euler_tour_rec(site p, int &position, int d) + { + if (pos.find(p) == pos.end()) + pos[p] = position; + + repr[position] = p; + depth[position] = d; + euler[position] = pos[p]; + ++position; + node& n = nodes(p); + + // typename p_array<mln_site(I)>::fwd_piter it(n.children); + // for_all(it) + // { + // build_euler_tour_rec(it.to_site(), position, d+1); + // depth[position] = d; // Not optimized + // euler[position] = pos[p]; + // repr[position] = p; // Pas necessaire? + // ++position; + // } + + for(unsigned i=0; i < n.children.size(); ++i) + { + build_euler_tour_rec(n.children[i], position, d+1); + depth[position] = d; // Not optimized + euler[position] = pos[p]; + repr[position] = p; // Pas necessaire? + ++position; + } + } + + + template <class I, class N> + void najman_wst<I, N>::build_euler_tour () + { + ctree_size = 0; + compute_ctree_size(Root); + + int size = 2 * ctree_size - 1; + + // FIXME : free this + euler = new int[size]; + depth = new int[size]; + repr = new site[size]; + + int position = 0; + build_euler_tour_rec(Root, position, 0); + } + + + template <class I, class N> + void najman_wst<I, N>::build_minim () + { + // compute_tree_size(); // Already done in build_euler_tour + int size = 2 * ctree_size - 1; + int logn = (int)(ceil(log((double)(size))/log(2.0))); + // minim.reserve(size); + minim = new int [logn * size]; // FIXME : Think about freeing this + Minim = new int* [logn]; + + Minim[0] = minim; + + // Init + for (int i = 0; i < size - 1; ++i) + if (depth[euler[i]] < depth[euler[i+1]]) { + Minim[0][i] = i; + } else { + Minim[0][i] = i+1; + } + + // Minim[0][size - 1] = size - 1; + + int k; + for (int j = 1; j < logn; j++) { + k = 1 << (j - 1); + Minim[j] = &minim[j * size]; + for (int i = 0; i < size; i++) { + if ((i + (k << 1)) >= size) { + //Minim[j][i] = size - 1; + } + else { + if (depth[euler[Minim[j - 1][i]]] <= depth[euler[Minim[j - 1][i + k]]]) + Minim[j][i] = Minim[j - 1][i]; + else + Minim[j][i] = Minim[j - 1][i + k]; + } + } + } + + } // void build_minim () + + + template <class I, class N> + typename najman_wst<I, N>::site najman_wst<I, N>::lca (site a, site b) + { + int + m = pos[a], + n = pos[b], + k; + + if (m == n) + return repr[m]; + + if (m > n) + { + k = n; + n = m; + m = k; + } + + k = (int)(log((double)(n - m))/log(2.)); + + if (depth[euler[Minim[k][m]]] < depth[euler[Minim[k][n - (1 << k)]]]) { + return repr[euler[Minim[k][m]]]; + } else { + return repr[euler[Minim[k][n - (1 << k)]]]; + } + } + + + template <class I, class N> + void najman_wst<I, N>::removeOneSonNodes(site *p, mln_ch_value(I, site) &newPar_node) + { + node &n = nodes(*p); + + if (n.children.size() == 1) // this node has 1 son, delete it + { + n.area = -1; + newPar_node(*p) = n.children[0]; + *p = n.children[0]; + removeOneSonNodes(p, newPar_node); + } + else // there is more than one son, recursive call + { + for (unsigned i = 0; i < n.children.size(); ++i) + removeOneSonNodes(&(n.children[i]), newPar_node); + } + } + + + template <class I, class N> + void najman_wst<I, N>::compressTree() + { + mln_ch_value(I, site) newPar_node(Par_node.domain(), Par_node.border()); + + // Remove the nodes with one son + removeOneSonNodes(&Root, newPar_node); + + // Update the references on deleted nodes + mln_piter(I) p(Par_node.domain()); + for_all(p) + while (nodes(Par_node(p)).area == -1) + Par_node(p) = newPar_node(Par_node(p)); + } + +# endif // MLN_INCLUDE_ONLY + + + }; // namespace morpho + +}; // namespace mln + +#endif // MLN_MORPHO_NAJMAN_WST_HH Index: abraham/mln/morpho/basic_najman.hh --- abraham/mln/morpho/basic_najman.hh (revision 2672) +++ abraham/mln/morpho/basic_najman.hh (working copy) @@ -33,8 +33,10 @@ #include <mln/estim/min_max.hh> #include <mln/math/sqr.hh> #include <mln/util/greater_psite.hh> +#include <mln/util/ord.hh> -#include <mln/core/site_set/p_priority_queue.hh> +#include <mln/core/site_set/p_priority.hh> +#include <mln/core/site_set/p_queue_fast.hh> #include <queue> #include <vector> @@ -78,8 +80,8 @@ } }; // struct node - I pima; - const Neighborhood<N>& nbh; + mln_exact(I) pima; + const mln_exact(N)& nbh; mln_ch_value(I, site) Par_node; mln_ch_value(I, site) Par_tree; @@ -148,6 +150,8 @@ compressTree(); build_euler_tour(); build_minim(); + m_watershed(); + w_watershed(); } @@ -269,22 +273,6 @@ return n; } - template <typename C> - void image_output(image2d<C>& img) - { - for(unsigned int i = 0; i < img.domain().len(0); ++i) - { - for(unsigned int j = 0; j < img.domain().len(1); ++j) - { - C l = (img(site(i, j)));//.row() * img.domain().len(1) + (img(site(i, j))).col(); - std::cout << " " << l << " "; - } - std::cout << std::endl; - } - } - - - void print_tree(site p) { node& n = nodes(p); @@ -335,12 +323,12 @@ } - site highest_fork (p_set<site>& components) + bool highest_fork (p_set<site>& components, site &r) { if (components.nsites() == 0) { std::cerr << "highest fork : empty set" << std::endl; - return site(-1, -1); + return false; } site @@ -360,12 +348,13 @@ } if (nodes(m).level == nodes(hfork).level) - return site(-1, -1); + return false; - return hfork; + r = hfork; + return true; } - site w_destructible(site p) + bool w_destructible(site p, site &r) { mln_niter(N) q(nbh, p); p_set<site> v; @@ -375,20 +364,45 @@ v.insert(Par_node(q)); if (v.nsites() == 0) - return site(-1, -1); + return false; - site hf = highest_fork(v); + site hf; + bool is_hf = highest_fork(v, hf); - if (hf == site(-1, -1)) - return min(v); + if (!is_hf) { + r = min(v); + return true; + } - if (nodes(hf).level <= pima(p)) - return hf; + if (nodes(hf).level <= pima(p)) { + r = hf; + return true; + } + return false; + } - return site(-1, -1); + bool highest_fork (p_set<site>& components) { + site i; + return highest_fork(components, i); } - site m_destructible(site p) + bool m_destructible(site p) { + site i; + return m_destructible(p, i); + } + + bool w_destructible(site p) { + site i; + return w_destructible(p, i); + } + + bool w_constructible(site p) { + site i; + return w_constructible(p, i); + } + + + bool m_destructible(site p, site &r) { mln_niter(N) q(nbh, p); p_set<site> v = p_set<site>(); @@ -398,21 +412,23 @@ v.insert(Par_node(q)); if (v.nsites() == 0) - return site(-1, -1); + return false; // if (nodes(min(v)).children.nsites() != 0) if (nodes(min(v)).children.size() != 0) - return (site(-1, -1)); + return false; - site hf = highest_fork(v); + bool is_hf = highest_fork(v); - if (hf == site(-1, -1)) - return min(v); + if (!is_hf) { + r = min(v); + return true; + } - return site(-1, -1); + return false; } - site w_constructible(site p) + bool w_constructible(site p, site &r) { mln_niter(N) q(nbh, p); p_set<site> v; @@ -422,10 +438,12 @@ v.insert(Par_node(q)); if (v.nsites() == 0) - return site(-1, -1); + return false; - if (v.nsites() == 1) - return v[0]; + if (v.nsites() == 1) { + r = v[0]; + return true; + } site c = max(v), @@ -445,9 +463,10 @@ } if (nodes(c).level <= pima(p)) - return site(-1, -1); + return false; - return c; + r = c; + return true; } void m_watershed () @@ -465,7 +484,7 @@ mln_piter(I) it(pima.domain()); for_all(it) - if (m_destructible(it.to_site()) != site(-1, -1)) + if (m_destructible(it.to_site())) { l.push(it.to_site()); isproc(it.to_site()) = true; @@ -481,9 +500,9 @@ // unmark p isproc(p) = false; - i = m_destructible(p); + bool is_m = m_destructible(p, i); - if (i != site(-1, -1)) + if (is_m) { pima(p) = nodes(i).level; Par_node(p) = i; @@ -491,7 +510,7 @@ mln_niter(N) q(nbh, p); for_all(q) if (pima.has(q) && !isproc(q)) - if (m_destructible(q) != site(-1, -1)) + if (m_destructible(q)) { // Add priority queue l.push(q); @@ -505,15 +524,9 @@ void w_watershed() { - std::map< mln_value(I), std::set<site> > L; + mln::p_priority< mln_value(I), p_queue_fast<site> > L; - // Setting the min and the max of the image - mln_value(I) k, kmin, kmax; - mln::estim::min_max(pima, kmin, kmax); - - // For k From kmin to kmax - 1 Do Lk <- <empty set> - for (k = kmin; k < kmax; k++) - L[k] = std::set<site>(); + mln_value(I) max = level::compute(accu::meta::max(), pima); // I K(pima.domain(), pima.border()); mln_ch_value(I, unsigned) K(pima.domain(), pima.border()); @@ -524,27 +537,26 @@ for_all(it) { site p = it.to_site(); + site i; // i <- W-Destructible(p) - site i = w_destructible(p); + bool is_w = w_destructible(p, i); // If i != infinity then - if (i != site(-1, -1)) + if (is_w) { - L[nodes(i).level].insert(p); + L.push(max - nodes(i).level, p); K(p) = nodes(i).level; H(p) = i; } } - for (k = kmin; k < kmax; k++) + while (!L.is_empty()) { - std::set<site>& Lk = L[k]; + mln_value(I) k = max - L.highest_priority(); - while (!Lk.empty()) - { - site p = *(Lk.begin()); - Lk.erase(p); + site p = L.front(); + L.pop(); if (K(p) == k) { @@ -555,13 +567,14 @@ for_all(q) if (pima.has(q) && k < pima(q)) { - site i = w_destructible(q); - if (i == site(-1, -1)) + site i; + bool is_w = w_destructible(q, i); + if (!is_w) K(q) = 10000; // FIXME : supposed to be infinity... else if (K(q) != nodes(i).level) { - L[nodes(i).level].insert(q); + L.push(max - nodes(i).level, q); K(q) = nodes(i).level; H(q) = i; } @@ -569,8 +582,6 @@ } } } - } - void revert_tree (site p) { @@ -628,7 +639,7 @@ // Mark enqueued sites mln_ch_value(I, bool) enqueued(pima.domain(), pima.border()); - p_priority_queue < site, mln_value(I) > l; + p_priority< mln_value(I), p_queue_fast<site> > l; // p_queue < site > m; std::queue<site> m; mln_value(I) max = mln_max(mln_value(I)); @@ -655,7 +666,7 @@ if (cmax.has(q) && !cmax(q) && !enqueued(q)) { enqueued(q) = true; - l.push(q, pima(q)); + l.push(pima(q), q); } m.pop(); } @@ -670,9 +681,10 @@ l.pop(); enqueued(x) = false; - site c = w_constructible(x); + site c; + bool is_w = w_constructible(x, c); - if (c != site(-1, -1)) + if (is_w) { pima(x) = nodes(c).level; Par_node(x) = c; @@ -692,7 +704,7 @@ { enqueued(q) = true; - l.push(q, pima(q)); + l.push(pima(q), q); } } // if (c != site(-1, -1)) } // while(!l.empty()) @@ -707,7 +719,7 @@ int *euler; int *depth; int ctree_size; - std::map<site, int> pos; + std::map<site, int, mln::util::ord<site> > pos; site *repr; int *minim; Index: abraham/mln/morpho/topo_wst.hh --- abraham/mln/morpho/topo_wst.hh (revision 0) +++ abraham/mln/morpho/topo_wst.hh (revision 0) @@ -0,0 +1,744 @@ +// Copyright (C) 2008 EPITA Research and Development Laboratory (LRDE) +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +#ifndef MLN_MORPHO_TOPO_WST_HH +# define MLN_MORPHO_TOPO_WST_HH + + + +#include <mln/level/sort_psites.hh> +#include <mln/level/fill.hh> +#include <mln/core/image/image2d.hh> +#include <mln/core/site_set/p_set.hh> +#include <mln/estim/min_max.hh> +#include <mln/math/sqr.hh> +#include <mln/util/greater_psite.hh> +#include <mln/util/ord.hh> +#include <mln/arith/revert.hh> + +#include <mln/core/site_set/p_priority.hh> +#include <mln/core/site_set/p_queue_fast.hh> + +#include <queue> +#include <vector> +#include <map> + +namespace mln +{ + namespace morpho + { + + template <class I, class N> + struct topo_wst + { + + // Typedefs + // -------- + typedef mln_site(I) site; + + // Component tree management + // ------------------------- + private: + struct node { + mln_value(I) level; + int area; + int highest; + typedef mln_site(I) site; + // Can't modify the sites in a p_array + // p_array<mln_site(I)> children; + std::vector<site> children; + + void addChildren(const node& n) + { + // typename p_array<mln_site(I)>::fwd_piter it(n.children); + // for (it.start(); + // it.is_valid(); + // it.next()) + // children.append(it.to_site()); + for (unsigned i=0; i < n.children.size(); ++i) + children.push_back(n.children[i]); + } + + void addChild(const site p) + { + // children.append(n); + children.push_back(p); + } + + mln_value(I) min() { return mln_min(mln_value(I)); } + mln_value(I) max() { return mln_max(mln_value(I)); } + + }; // struct node + + const mln_exact(N)& nbh; + mln_ch_value(I, site) Par_node; + mln_ch_value(I, site) Par_tree; + + mln_ch_value(I, int) Rnk_tree; + mln_ch_value(I, int) Rnk_node; + mln_ch_value(I, site) subtreeRoot; + mln_ch_value(I, node) nodes; + site Root; + + void MakeSet_tree(site x); + void MakeSet_node(site x); + site Find_tree(site x); + void swap(site& x, site& y); + site Find_node(site x); + void BuildComponentTree(); + site MergeNode(site& node1, site& node2); + site Link_tree(site x, site y); + site Link_node(site x, site y); + node MakeNode(int level); + + + // Watershed algorithm + // ------------------- + + private: + mln_ch_value(I, bool) isproc; + + // Ctor + public: + topo_wst(const Image& i, + const Neighborhood<N>& n); + + + // Result + public: + mln_exact(I) pima; + + public: + void go(); + + private: + site min (p_set<site>& components); + site max (p_set<site>& components); + bool highest_fork (p_set<site>& components); + bool highest_fork (p_set<site>& components, site &r); + bool w_constructible(site p, site &r); + bool w_constructible(site p); + void topo_watershed(); + + + // Optimized LCA Algorithm + private: + int *euler; + int *depth; + int ctree_size; + std::map<site, int, mln::util::ord<site> > pos; + site *repr; + + int *minim; + int **Minim; + + void compute_ctree_size (site p); + void build_euler_tour_rec(site p, int &position, int d); + void build_euler_tour(); + void build_minim (); + site lca (site a, site b); + void removeOneSonNodes(site *p, mln_ch_value(I, site) &newPar_node); + void compressTree(); + }; // struct topo_wst + +# ifndef MLN_INCLUDE_ONLY + + // Ctor + template <class I, class N> + topo_wst<I, N>::topo_wst(const Image& i, + const Neighborhood<N>& n) + : pima(exact(i)), + nbh(exact(n)), + Par_node(exact(i).domain(), exact(i).border()), + Par_tree(exact(i).domain(), exact(i).border()), + Rnk_tree(exact(i).domain(), exact(i).border()), + Rnk_node(exact(i).domain(), exact(i).border()), + subtreeRoot(exact(i).domain(), exact(i).border()), + nodes(exact(i).domain(), exact(i).border()), + isproc(exact(i).domain(), exact(i).border()) + { + } + + template <class I, class N> + void topo_wst<I, N>::MakeSet_tree(site x) + { + Par_tree(x) = x; + Rnk_tree(x) = 0; + } + + template <class I, class N> + void topo_wst<I, N>::MakeSet_node(site x) + { + Par_node(x) = x; + Rnk_node(x) = 0; + } + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::Find_tree(site x) + { + if (Par_tree(x) != x) + Par_tree(x) = Find_tree(Par_tree(x)); + return Par_tree(x); + } + + template <class I, class N> + void topo_wst<I, N>::swap(site& x, site& y) + { + site memo = x; + x = y; + y = memo; + } + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::Find_node(site x) + { + if (Par_node(x) != x) + Par_node(x) = Find_node(Par_node(x)); + return Par_node(x); + } + + template <class I, class N> + void topo_wst<I, N>::go() + { + BuildComponentTree(); + compressTree(); + arith::revert_inplace(pima); + arith::revert_inplace(nodes); + build_euler_tour(); + build_minim(); + topo_watershed(); + arith::revert_inplace(pima); + } + + template <class I, class N> + void topo_wst<I, N>::BuildComponentTree() + { + // Init: + + // Sort the sites in increasing order + p_array<mln_site(I)> S; + S = level::sort_psites_increasing(pima); + + // Clear the marker map + level::fill(isproc, false); + for (int ip = 0; ip < int(S.nsites()); ++ip) + { + site p = S[ip]; + MakeSet_node(p); + MakeSet_tree(p); + // if (subtreeRoot.hold(p)) + subtreeRoot(p) = p; + // if (nodes.hold(p)) + nodes(p) = MakeNode(pima(p)); + } + + + + typename p_array<mln_site(I)>::fwd_piter ip(S); + for_all(ip) + { + site p = ip.to_site(); + + site curCanonicalElt = Find_tree(p); + site curNode = Find_node(subtreeRoot(curCanonicalElt)); + + mln_niter(N) q(nbh, ip); + for_all(q) + if (pima.has(q) and isproc(q) and pima(q) <= pima(p)) + { + site adjCanonicalElt = Find_tree(q); + site adjNode = Find_node(subtreeRoot(adjCanonicalElt)); + if (curNode != adjNode) + { + if (nodes(curNode).level == nodes(adjNode).level) + curNode = MergeNode(adjNode, curNode); + else + { + nodes(curNode).addChild(adjNode); + nodes(curNode).area += nodes(adjNode).area; + nodes(curNode).highest += nodes(adjNode).highest; + } + } + + curCanonicalElt = Link_tree(adjCanonicalElt, curCanonicalElt); + subtreeRoot(curCanonicalElt) = curNode; + } + isproc(p) = true; + } + // Pour garder une map de correspondance site <-> local_root + // for (int ip = 0; ip < int(S.size()); ++ip) + // { + // site p = S[ip]; + // M(p) = Find_node(p); + // } + + mln_piter(I) r(Par_node.domain()); + for_all(r) + Par_node(r) = Find_node(r); + + Root = subtreeRoot(Find_tree(Find_node(site(0, 0)))); + } + + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::MergeNode(site& node1, site& node2) + { + site tmpNode = Link_node(node1, node2); + site tmpNode2; + if (tmpNode == node2) + { + nodes(node2).addChildren(nodes(node1)); + tmpNode2 = node1; + } + else + { + nodes(node1).addChildren(nodes(node2)); + tmpNode2 = node2; + } + nodes(tmpNode).area += nodes(tmpNode2).area; + nodes(tmpNode).highest += nodes(tmpNode2).highest; + return tmpNode; + } + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::Link_tree(site x, site y) + { + if (Rnk_tree(x) > Rnk_tree(y)) + swap(x, y); + else + if (Rnk_tree(x) == Rnk_tree(y)) + Rnk_tree(y) += 1; + Par_tree(x) = y; + return y; + } + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::Link_node(site x, site y) + { + if (Rnk_node(x) > Rnk_node(y)) + swap(x, y); + else + if (Rnk_node(x) == Rnk_node(y)) + Rnk_node(y) += 1; + Par_node(x) = y; + return y; + } + + template <class I, class N> + typename topo_wst<I, N>::node topo_wst<I, N>::MakeNode(int level) + { + node n; + n.level = level; + n.area = 1; + n.highest = level; + return n; + } + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::min (p_set<site>& components) + { + if (components.nsites() == 0) + return site(-1, -1); + + typename p_set<site>::fwd_piter it(components); + site min = components[0]; + + for_all(it) + if (pima(it.to_site()) < pima(min)) + min = it.to_site(); + + return min; + } + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::max (p_set<site>& components) + { + if (components.nsites() == 0) + return site(-1, -1); + + typename p_set<site>::fwd_piter it(components); + site max = components[0]; + + for_all(it) + if (pima(it.to_site()) > pima(max)) + max = it.to_site(); + + return max; + } + + + template <class I, class N> + bool topo_wst<I, N>::highest_fork (p_set<site>& components, site &r) + { + if (components.nsites() == 0) + { + std::cerr << "highest fork : empty set" << std::endl; + return false; + } + + site + m = min(components), + hfork = m; + + typename p_set<site>::fwd_piter it(components); + for_all(it) + { + // Can't remove the site from the set + if (it.to_site() == m) + continue; + + site c = lca(hfork, it.to_site()); + if (c != it.to_site()) + hfork = c; + } + + if (nodes(m).level == nodes(hfork).level) + return false; + + r = hfork; + return true; + } + + template <class I, class N> + bool topo_wst<I, N>::highest_fork (p_set<site>& components) { + site i; + return highest_fork(components, i); + } + + template <class I, class N> + bool topo_wst<I, N>::w_constructible(site p) { + site i; + return w_constructible(p, i); + } + + template <class I, class N> + bool topo_wst<I, N>::w_constructible(site p, site &r) + { + mln_niter(N) q(nbh, p); + p_set<site> v; + + for_all(q) + if (pima.has(q) && pima(q) > pima(p)) + v.insert(Par_node(q)); + + if (v.nsites() == 0) + return false; + + if (v.nsites() == 1) { + r = v[0]; + return true; + } + + site + c = max(v), + cmax = c; + + typename p_set<site>::fwd_piter it(v); + for_all(it) + { + // Can't remove the site from the set + if (it.to_site() == cmax) + continue; + + site c1 = lca(c, it.to_site()); + + if (c1 != it.to_site()) + c = c1; + } + + if (nodes(c).level <= pima(p)) + return false; + + r = c; + return true; + } + + + template <class I, class N> + void topo_wst<I, N>::topo_watershed() + { + // Maxima components + mln_ch_value(I, bool) cmax(pima.domain(), pima.border()); + + // Mark enqueued sites + mln_ch_value(I, bool) enqueued(pima.domain(), pima.border()); + + p_priority< mln_value(I), p_queue_fast<site> > l; + // p_queue < site > m; + std::queue<site> m; + mln_value(I) max = mln_max(mln_value(I)); + + std::cout << "Init" << std::endl; + + // Flag C-maxima + level::fill(cmax, false); + level::fill(enqueued, false); + + mln_piter(I) it(Par_node.domain()); + for_all(it) + // if (nodes(Par_node(it.to_site())).children.nsites() == 0) + if (nodes(Par_node(it.to_site())).children.size() == 0) + { + cmax(it.to_site()) = true; + m.push(it.to_site()); + } + + while (!m.empty()) + { + mln_niter(N) q(nbh, m.front()); + for_all(q) + if (cmax.has(q) && !cmax(q) && !enqueued(q)) + { + enqueued(q) = true; + l.push(pima(q), q); + } + m.pop(); + } + + + std::cout << "end" << std::endl; + + // Main loop + while(!l.is_empty()) + { + site x = l.front(); + l.pop(); + enqueued(x) = false; + + site c; + bool is_w = w_constructible(x, c); + + if (is_w) + { + pima(x) = nodes(c).level; + Par_node(x) = c; + + // if (nodes(c).children.nsites() == 0) + if (nodes(c).children.size() == 0) + cmax(x) = true; + else + // if (nodes(c).children.nsites() > 1) + if (nodes(c).children.size() == 1) + std::cerr << "ERREUR COMPOSANTE BRANCHE " << nodes(c).children.size() << std::endl; + + + mln_niter(N) q(nbh, x); + for_all(q) + if (pima.has(q) && !cmax(q) && !enqueued(q)) + { + enqueued(q) = true; + + l.push(pima(q), q); + } + } // if (c != site(-1, -1)) + } // while(!l.empty()) + + for_all(it) + pima(it.to_site()) = nodes(Par_node(it.to_site())).level; + } + + + + template <class I, class N> + void topo_wst<I, N>::compute_ctree_size (site p) + { + ctree_size += 1; + node& n = nodes(p); + + // typename p_array<mln_site(I)>::fwd_piter it(n.children); + // for_all(it) + // compute_ctree_size(it.to_site()); + + for (unsigned i=0; i < n.children.size(); ++i) + compute_ctree_size(n.children[i]); + } + + + template <class I, class N> + void topo_wst<I, N>::build_euler_tour_rec(site p, int &position, int d) + { + if (pos.find(p) == pos.end()) + pos[p] = position; + + repr[position] = p; + depth[position] = d; + euler[position] = pos[p]; + ++position; + node& n = nodes(p); + + // typename p_array<mln_site(I)>::fwd_piter it(n.children); + // for_all(it) + // { + // build_euler_tour_rec(it.to_site(), position, d+1); + // depth[position] = d; // Not optimized + // euler[position] = pos[p]; + // repr[position] = p; // Pas necessaire? + // ++position; + // } + + for(unsigned i=0; i < n.children.size(); ++i) + { + build_euler_tour_rec(n.children[i], position, d+1); + depth[position] = d; // Not optimized + euler[position] = pos[p]; + repr[position] = p; // Pas necessaire? + ++position; + } + } + + + template <class I, class N> + void topo_wst<I, N>::build_euler_tour () + { + ctree_size = 0; + compute_ctree_size(Root); + + int size = 2 * ctree_size - 1; + + // FIXME : free this + euler = new int[size]; + depth = new int[size]; + repr = new site[size]; + + int position = 0; + build_euler_tour_rec(Root, position, 0); + } + + + template <class I, class N> + void topo_wst<I, N>::build_minim () + { + // compute_tree_size(); // Already done in build_euler_tour + int size = 2 * ctree_size - 1; + int logn = (int)(ceil(log((double)(size))/log(2.0))); + // minim.reserve(size); + minim = new int [logn * size]; // FIXME : Think about freeing this + Minim = new int* [logn]; + + Minim[0] = minim; + + // Init + for (int i = 0; i < size - 1; ++i) + if (depth[euler[i]] < depth[euler[i+1]]) { + Minim[0][i] = i; + } else { + Minim[0][i] = i+1; + } + + // Minim[0][size - 1] = size - 1; + + int k; + for (int j = 1; j < logn; j++) { + k = 1 << (j - 1); + Minim[j] = &minim[j * size]; + for (int i = 0; i < size; i++) { + if ((i + (k << 1)) >= size) { + //Minim[j][i] = size - 1; + } + else { + if (depth[euler[Minim[j - 1][i]]] <= depth[euler[Minim[j - 1][i + k]]]) + Minim[j][i] = Minim[j - 1][i]; + else + Minim[j][i] = Minim[j - 1][i + k]; + } + } + } + + } // void build_minim () + + + template <class I, class N> + typename topo_wst<I, N>::site topo_wst<I, N>::lca (site a, site b) + { + int + m = pos[a], + n = pos[b], + k; + + if (m == n) + return repr[m]; + + if (m > n) + { + k = n; + n = m; + m = k; + } + + k = (int)(log((double)(n - m))/log(2.)); + + if (depth[euler[Minim[k][m]]] < depth[euler[Minim[k][n - (1 << k)]]]) { + return repr[euler[Minim[k][m]]]; + } else { + return repr[euler[Minim[k][n - (1 << k)]]]; + } + } + + + template <class I, class N> + void topo_wst<I, N>::removeOneSonNodes(site *p, mln_ch_value(I, site) &newPar_node) + { + node &n = nodes(*p); + + if (n.children.size() == 1) // this node has 1 son, delete it + { + n.area = -1; + newPar_node(*p) = n.children[0]; + *p = n.children[0]; + removeOneSonNodes(p, newPar_node); + } + else // there is more than one son, recursive call + { + for (unsigned i = 0; i < n.children.size(); ++i) + removeOneSonNodes(&(n.children[i]), newPar_node); + } + } + + + template <class I, class N> + void topo_wst<I, N>::compressTree() + { + mln_ch_value(I, site) newPar_node(Par_node.domain(), Par_node.border()); + + // Remove the nodes with one son + removeOneSonNodes(&Root, newPar_node); + + // Update the references on deleted nodes + mln_piter(I) p(Par_node.domain()); + for_all(p) + while (nodes(Par_node(p)).area == -1) + Par_node(p) = newPar_node(Par_node(p)); + } + +# endif // MLN_INCLUDE_ONLY + + + }; // namespace morpho + +}; // namespace mln + +#endif // MLN_MORPHO_TOPO_WST_HH Index: abraham/mln/core/image/thru.hh --- abraham/mln/core/image/thru.hh (revision 2672) +++ abraham/mln/core/image/thru.hh (working copy) @@ -36,7 +36,9 @@ */ # include <mln/core/internal/image_value_morpher.hh> +# include <mln/trait/images.hh> # include <mln/value/set.hh> +# include <mln/core/image/shell.hh> namespace mln { @@ -50,8 +52,8 @@ template <typename F, typename I> struct data< thru<F,I> > { - data(const I& ima); - const I& ima_; + data( I& ima); + I& ima_; }; } // end of namespace mln::internal @@ -64,7 +66,9 @@ template <typename F, typename I> struct image_< thru<F,I> > : default_image_morpher< I, mln_result(F), thru<F,I> > { - typedef trait::image::value_io::read_only value_io; + typedef trait::image::value_io::read_write value_io; + typedef trait::image::value_access::computed value_access; + typedef trait::image::value_storage::disrupted value_storage; }; } // end of namespace mln::trait @@ -85,28 +89,22 @@ typedef mln_result(F) rvalue; /// Return type of read-write access. - typedef mln_result(F) lvalue; + typedef shell<F,I> lvalue; /// Skeleton. typedef thru< tag::value_<mln_result(F)>, tag::image_ > skeleton; /// Constructor. - thru(const Image& ima); + thru(Image& ima); /// Initialize an empty image. - void init_(const Image& ima); + void init_(Image& ima); /// Read-only access of pixel value at point site \p p. mln_result(F) operator()(const mln_psite(I)& p) const; - /// Mutable access is only OK for reading (not writing). - // T operator()(const mln_psite(I)& p); - }; - - template <typename F, typename I> - struct thru <F, const I> - { - mln_result(F) operator()(const mln_psite(I)& p); + /// Mutable access is for reading and writing. + lvalue operator()(const mln_psite(I)& p); }; # ifndef MLN_INCLUDE_ONLY @@ -119,7 +117,7 @@ template <typename F, typename I> inline - data< thru<F,I> >::data(const I& ima) + data< thru<F,I> >::data(I& ima) : ima_(ima) { } @@ -130,22 +128,22 @@ template <typename F, typename I> inline - thru<F,I>::thru(const Image& ima) + thru<F,I>::thru(Image& ima) { mln_precondition(exact(ima).has_data()); this->data_ = new internal::data< thru<F,I> >(exact(ima)); } - /* + template <typename F, typename I> inline void - thru<F,I>::init_(const Image& ima) + thru<F,I>::init_(Image& ima) { mln_precondition(exact(ima).has_data()); this->data_ = new internal::data<thru<F,I> >(exact(ima)); } - */ + template <typename F, typename I> inline @@ -158,10 +156,10 @@ template <typename F, typename I> inline - mln_result(F) + shell<F, I> thru<F,I>::operator()(const mln_psite(I)& p) { - return *(F*)(void*)&( this->data_->ima_(p) ); + return shell<F, I>( this->data_->ima_, p ); } # endif // ! MLN_INCLUDE_ONLY Index: abraham/mln/core/image/shell.hh --- abraham/mln/core/image/shell.hh (revision 0) +++ abraham/mln/core/image/shell.hh (revision 0) @@ -0,0 +1,118 @@ +// Copyright (C) 2007 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +#ifndef MLN_CORE_VALUE_SHELL_HH +# define MLN_CORE_VALUE_SHELL_HH + +# include <mln/core/concept/image.hh> + +namespace mln +{ + + // FIXME : traits + + template <typename F, typename I, class C> + struct shell_write; + + template <typename F, typename I> + struct shell_write<F, I, Function_v2w2v<void> > + { + mln_value(I) set_ (I &ima, const mln_site(I) &s, const typename F::result &v); + }; + + template <typename F, typename I> + struct shell_write<F, I, Function_v2w_w2v<void> > + { + mln_value(I) set_ (I &ima, const mln_site(I) &s, const typename F::result &v); + }; + + + template <typename F, typename I> + struct shell; + + template <typename F, typename I> + struct shell : shell_write<F, I, typename F::category> // FIXME : inherit of value_base or sth? + { + typedef mln_value(I) value; + + // Ctor + shell(Image &ima, const mln_site(I) &s); + + // Read + operator value (); + + // Write + mln_value(I) operator= (typename F::result); + + protected : + I &ima; + mln_site(I) s; + }; + +# ifndef MLN_INCLUDE_ONLY + + // Ctor + + template <typename F, typename I> + shell<F, I>::shell(Image &ima, const mln_site(I) &s) + : ima(exact(ima)), s(s) + { } + + + // Read for everyone + template <typename F, typename I> + shell<F, I>::operator value () + { + return F()(ima(s)); + } + + // Write for everyone + template <typename F, typename I> + mln_value(I) shell<F, I>::operator= (typename F::result v) + { + set_(ima, s, v); + } + + template <typename F, typename I> + mln_value(I) shell_write<F, I, Function_v2w2v<void> >::set_ (I &ima, const mln_site(I) &s, const typename F::result &v) + { + ima(s) = F().f_1(v); + return ima(s); + } + + template <typename F, typename I> + mln_value(I) shell_write<F, I, Function_v2w_w2v<void> >::set_ (I &ima, const mln_site(I) &s, const typename F::result &v) + { + ima(s) = F().f_1(ima(s), v); + return ima(s); + } + +# endif // MLN_INCLUDE_ONLY + +}; // end of namespace mln + +#endif // MLN_CORE_VALUE_SHELL_HH Index: abraham/mln/core/concept/function.hh --- abraham/mln/core/concept/function.hh (revision 2672) +++ abraham/mln/core/concept/function.hh (working copy) @@ -160,7 +160,7 @@ template <typename E> struct Function_v2w_w2v : public Function<E> { - typedef Function_v2w2v<void> category; + typedef Function_v2w_w2v<void> category; /* result operator() (value); Index: abraham/mln/core/concept/meta_fun.hh --- abraham/mln/core/concept/meta_fun.hh (revision 0) +++ abraham/mln/core/concept/meta_fun.hh (revision 0) @@ -0,0 +1,83 @@ +// Copyright (C) 2007 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +#ifndef MLN_CORE_CONCEPT_META_FUN_HH +# define MLN_CORE_CONCEPT_META_FUN_HH + +/*! \file mln/core/concept/meta_fun.hh + * + * \brief Definition of concept of meta function. + */ + +# include <mln/core/concept/object.hh> +# include <mln/core/concept/function.hh> + +namespace mln +{ + + template <class M, class T> + struct function; + + namespace meta + { + + template <class M> + struct impl + { + + template <class T> + struct info + { + typedef function<M, T> F; + typedef typename F::result result; + typedef typename F::lresult lresult; + }; + + template <class T> + typename info<T>::result + operator()(const T& t) const + { + function<M,T> f; + return f.read(t); + } + + template <class T> + T& + f_1(typename info<T>::result v, T& t) + { + function<M,T> f; + f.write(t) = v; + return t; + } + + }; + + } // end of namespace meta + +} // end of namespace mln + +#endif // MLN_CORE_CONCEPT_META_FUN_HH Index: abraham/mln/fun/meta/red.hh --- abraham/mln/fun/meta/red.hh (revision 0) +++ abraham/mln/fun/meta/red.hh (revision 0) @@ -0,0 +1,45 @@ +// Copyright (C) 2007 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. + +#ifndef MLN_FUN_META_RED_HH +# define MLN_FUN_META_RED_HH + +# include <mln/core/concept/meta_fun.hh> + +namespace mln { + + namespace meta { + + struct red : impl<red> { }; + + }; + + meta::red red; // fun obj + +}; + +#endif // MLN_FUN_META_RED_HH Index: abraham/mln/fun/v2w_w2v/norm.hh --- abraham/mln/fun/v2w_w2v/norm.hh (revision 2672) +++ abraham/mln/fun/v2w_w2v/norm.hh (working copy) @@ -71,7 +71,7 @@ * \see mln::norm::l2. */ template <typename V, typename R> - struct l2_norm : public Function_v2v< l2_norm<V, R> > + struct l2_norm : public Function_v2w_w2v< l2_norm<V, R> > { typedef R result; R operator()(const V& v) const; @@ -85,7 +85,7 @@ * \see mln::norm::linfty. */ template <typename V, typename R> - struct linfty_norm : public Function_v2v< linfty_norm<V, R> > + struct linfty_norm : public Function_v2w_w2v< linfty_norm<V, R> > { typedef R result; R operator()(const V& v) const; Index: vigouroux/tests.cc --- vigouroux/tests.cc (revision 2672) +++ vigouroux/tests.cc (working copy) @@ -1,4 +1,4 @@ -#include <mln/core/image_if_value.hh> +#include <mln/core/image/image_if.hh> #include <mln/core/alias/w_window2d_int.hh> #include <mln/display/show.hh> Index: vigouroux/color/rgb_to_cmy.hh --- vigouroux/color/rgb_to_cmy.hh (revision 2672) +++ vigouroux/color/rgb_to_cmy.hh (working copy) @@ -1,4 +1,4 @@ -#include <mln/core/image_if_value.hh> +#include <mln/core/image/image_if.hh> #include <mln/core/alias/w_window2d_int.hh> #include <mln/display/show.hh> Index: vigouroux/color/rgb_to_xyz.hh --- vigouroux/color/rgb_to_xyz.hh (revision 2672) +++ vigouroux/color/rgb_to_xyz.hh (working copy) @@ -1,7 +1,7 @@ #ifndef OLENA_CONVERT_RGBXYZ_HH # define OLENA_CONVERT_RGBXYZ_HH -# include <mln/core/image_if_value.hh> +# include <mln/core/image/image_if.hh> # include <mln/core/alias/w_window2d_int.hh> # include <mln/display/show.hh> Index: vigouroux/color/rgb_to_hsi.hh --- vigouroux/color/rgb_to_hsi.hh (revision 2672) +++ vigouroux/color/rgb_to_hsi.hh (working copy) @@ -1,7 +1,7 @@ #include <cmath> -#include <mln/core/image_if_value.hh> +#include <mln/core/image/image_if.hh> #include <mln/core/alias/w_window2d_int.hh> #include <mln/display/show.hh> Index: vigouroux/color/rgb_to_yuv.hh --- vigouroux/color/rgb_to_yuv.hh (revision 2672) +++ vigouroux/color/rgb_to_yuv.hh (working copy) @@ -1,4 +1,4 @@ -#include <mln/core/image_if_value.hh> +#include <mln/core/image/image_if.hh> #include <mln/core/alias/w_window2d_int.hh> #include <mln/display/show.hh>

1 0

cleanup-2008 2672: De-activate not compiling tests in level transform.
by Thierry Geraud 24 Oct '08

24 Oct '08

https://svn.lrde.epita.fr/svn/oln/branches/cleanup-2008/milena Index: ChangeLog from Thierry Geraud <thierry.geraud(a)lrde.epita.fr> De-activate not compiling tests in level transform. * doc/examples/tuto_bis.cc: Layout. (label): Set no border. (lab): Change extension from pw fun to image. That fixes some weird behavior. * mln/level/transform.hh (todo): New. De-activate tests since Nicolas has not fixed it yet. * mln/level/stretch.hh (FIXME): Fix. Avoid escaping allows for properly tracing. * mln/estim/nsites.hh: Rename as... * mln/geom/nsites.hh: ...this. Update. * mln/level/sort_psites.hh: Use geom::nsites. Remember that only some images have the nsites method. * mln/trace/exiting.hh: Print the scope when exiting. It eases debugging (reading traces). * mln/morpho/closing_volume.hh,. * mln/morpho/meyer_wst.hh: Add trace equipment doc/examples/tuto_bis.cc | 11 +++++++---- mln/geom/nsites.hh | 34 +++++++++++++++++----------------- mln/level/sort_psites.hh | 8 +++++--- mln/level/stretch.hh | 6 +++--- mln/level/transform.hh | 18 +++++++++++------- mln/morpho/closing_volume.hh | 2 ++ mln/morpho/meyer_wst.hh | 2 ++ mln/trace/exiting.hh | 6 +++--- 8 files changed, 50 insertions(+), 37 deletions(-) Index: doc/examples/tuto_bis.cc --- doc/examples/tuto_bis.cc (revision 2671) +++ doc/examples/tuto_bis.cc (working copy) @@ -20,6 +20,7 @@ # include <mln/level/paste.hh> # include <mln/level/fill.hh> # include <mln/level/transform.hh> +# include <mln/extension/fill.hh> # include <mln/morpho/meyer_wst.hh> @@ -113,7 +114,7 @@ dilation(const I& input, const N& nbh) { typedef mln_value(I) V; - // extension::fill(input, mln_min(V)); + // FIXME: extension::fill(input, mln_min(V)); mln_concrete(I) output; initialize(output, input); @@ -133,10 +134,12 @@ } // mln::morpho + } // mln + // Functions inline @@ -239,8 +242,7 @@ // 1 1 - image2d<unsigned> label(ima.bbox(), 1); - border::fill(label, 0); + image2d<unsigned> label(ima.bbox(), 0); level::fill(label, 9); debug::println(label); // 9 9 9 9 9 @@ -309,7 +311,8 @@ // // 9 9 9 - level::paste(morpho::dilation(extend(lab, pw::value(label)), + + level::paste(morpho::dilation(extend(lab, label), c4()), label); Index: mln/level/transform.hh --- mln/level/transform.hh (revision 2671) +++ mln/level/transform.hh (working copy) @@ -1,4 +1,4 @@ -// Copyright (C) 2007 EPITA Research and Development Laboratory +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory // // This file is part of the Olena Library. This library is free // software; you can redistribute it and/or modify it under the terms @@ -33,6 +33,8 @@ * \brief Transform the contents of an image into another one. * * \todo Clean this file + overload with pixel iterators. + * + * \todo Re-activate tests and make them static. */ # include <mln/core/concept/image.hh> @@ -91,10 +93,11 @@ (void) output; // Properties check - mln_precondition((mlc_is(mln_trait_image_pw_io(O), - trait::image::pw_io::read_write)::value || - mlc_is(mln_trait_image_vw_io(O), - trait::image::vw_io::read_write)::value)); + // FIXME: Re-activate! +// mln_precondition((mlc_is(mln_trait_image_pw_io(O), +// trait::image::pw_io::read_write)::value || +// mlc_is(mln_trait_image_vw_io(O), +// trait::image::vw_io::read_write)::value)); // FIXME Convert test mlc_converts_to(mln_result(F), mln_value(O))::check(); @@ -129,8 +132,9 @@ level::internal::transform_tests(input, f, output); - mlc_is(mln_trait_image_pw_io(O), - trait::image::pw_io::read_write)::check(); + // FIXME: Re-activate! +// mlc_is(mln_trait_image_pw_io(O), +// trait::image::pw_io::read_write)::check(); mln_piter(I) p(input.domain()); for_all(p) Index: mln/level/stretch.hh --- mln/level/stretch.hh (revision 2671) +++ mln/level/stretch.hh (working copy) @@ -72,8 +72,8 @@ mln_value(I) min_, max_; estim::min_max(input, min_, max_); - if (max_ == min_) - return; // FIXME + if (max_ != min_) + { float min = float(min_), max = float(max_); const float epsilon = mln_epsilon(float); float m = 0.0f - 0.5f + epsilon; @@ -82,7 +82,7 @@ float b = (m * max - M * min) / (max - min); fun::v2v::linear<float, float, int> f(a, b); level::transform(input, f, output); - + } trace::exiting("level::impl::stretch"); } Index: mln/level/sort_psites.hh --- mln/level/sort_psites.hh (revision 2671) +++ mln/level/sort_psites.hh (working copy) @@ -1,4 +1,5 @@ -// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory (LRDE) +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory +// (LRDE) // // This file is part of the Olena Library. This library is free // software; you can redistribute it and/or modify it under the terms @@ -39,6 +40,7 @@ # include <mln/convert/to_p_array.hh> # include <mln/histo/compute.hh> # include <mln/util/ord.hh> +# include <mln/geom/nsites.hh> namespace mln @@ -154,7 +156,7 @@ loc[i] = loc[i-1] + h[i-1]; // computing output data - std::vector<mln_psite(I)> vec(input.nsites()); + std::vector<mln_psite(I)> vec(geom::nsites(input)); mln_fwd_piter(I) p(input.domain()); for_all(p) vec[loc[vset.index_of(input(p))]++] = p; @@ -198,7 +200,7 @@ loc[i] = loc[i+1] + h[i+1]; // computing output data - std::vector<mln_psite(I)> vec(input.nsites()); + std::vector<mln_psite(I)> vec(geom::nsites(input)); mln_fwd_piter(I) p(input.domain()); for_all(p) vec[loc[vset.index_of(input(p))]++] = p; Index: mln/geom/nsites.hh --- mln/geom/nsites.hh (revision 2670) +++ mln/geom/nsites.hh (working copy) @@ -25,10 +25,10 @@ // reasons why the executable file might be covered by the GNU General // Public License. -#ifndef MLN_ESTIM_NSITES_HH -# define MLN_ESTIM_NSITES_HH +#ifndef MLN_GEOM_NSITES_HH +# define MLN_GEOM_NSITES_HH -/*! \file mln/estim/nsites.hh +/*! \file mln/geom/nsites.hh * * \brief Compute the number of sites of an image or a site set. */ @@ -40,7 +40,7 @@ namespace mln { - namespace estim + namespace geom { /// Compute the number of sites of the site set \p input. @@ -69,7 +69,7 @@ template <typename S> unsigned nsites(const Site_Set<S>& s_) { - trace::entering("estim::impl::generic::nsites"); + trace::entering("geom::impl::generic::nsites"); const S& s = exact(s_); mln_precondition(s.is_valid()); @@ -78,11 +78,11 @@ for_all(p) ++n; - trace::exiting("estim::impl::generic::nsites"); + trace::exiting("geom::impl::generic::nsites"); return n; } - } // end of namespace mln::estim::impl::generic + } // end of namespace mln::geom::impl::generic // A single specialization. @@ -91,13 +91,13 @@ inline unsigned nsites_method(const Site_Set<S>& s) { - trace::entering("estim::impl::nsites_method"); + trace::entering("geom::impl::nsites_method"); unsigned n = exact(s).nsites(); - trace::exiting("estim::impl::nsites_method"); + trace::exiting("geom::impl::nsites_method"); return n; } - } // end of namespace mln::estim::impl + } // end of namespace mln::geom::impl @@ -132,7 +132,7 @@ s); } - } // end of namespace mln::estim::internal + } // end of namespace mln::geom::internal @@ -142,12 +142,12 @@ inline unsigned nsites(const Site_Set<S>& s) { - trace::entering("estim::nsites"); + trace::entering("geom::nsites"); mln_precondition(exact(s).is_valid()); unsigned n = internal::nsites_dispatch(s); - trace::exiting("estim::nsites"); + trace::exiting("geom::nsites"); return n; } @@ -155,7 +155,7 @@ inline unsigned nsites(const Image& input_) { - trace::entering("estim::nsites"); + trace::entering("geom::nsites"); const I& input = exact(input_); mln_precondition(input.has_data()); @@ -164,15 +164,15 @@ // Relies on the nsites routines on a site set. unsigned n = internal::nsites_dispatch(input.domain()); - trace::exiting("estim::nsites"); + trace::exiting("geom::nsites"); return n; } # endif // ! MLN_INCLUDE_ONLY - } // end of namespace mln::estim + } // end of namespace mln::geom } // end of namespace mln -#endif // ! MLN_ESTIM_NSITES_HH +#endif // ! MLN_GEOM_NSITES_HH Index: mln/trace/exiting.hh --- mln/trace/exiting.hh (revision 2671) +++ mln/trace/exiting.hh (working copy) @@ -1,4 +1,4 @@ -// Copyright (C) 2007 EPITA Research and Development Laboratory +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory // // This file is part of the Olena Library. This library is free // software; you can redistribute it and/or modify it under the terms @@ -50,14 +50,14 @@ # ifndef MLN_INCLUDE_ONLY inline - void exiting(const std::string&) + void exiting(const std::string& str) { if (quiet) return; --tab; for (unsigned i = 0; i < tab; ++i) std::cout << " "; - std::cout << "}" << std::endl; + std::cout << "} " << str << std::endl; } # endif // ! MLN_INCLUDE_ONLY Index: mln/morpho/closing_volume.hh --- mln/morpho/closing_volume.hh (revision 2671) +++ mln/morpho/closing_volume.hh (working copy) @@ -54,9 +54,11 @@ void closing_volume(const Image& input, const Neighborhood<N>& nbh, std::size_t lambda, Image<O>& output) { + trace::entering("morpho::closing_volume"); mln_precondition(exact(output).domain() == exact(input).domain()); // FIXME: Change sig of closing_attribute! closing_attribute< accu::volume >(input, nbh, lambda, output); + trace::exiting("morpho::closing_volume"); } # endif // ! MLN_INCLUDE_ONLY Index: mln/morpho/meyer_wst.hh --- mln/morpho/meyer_wst.hh (revision 2671) +++ mln/morpho/meyer_wst.hh (working copy) @@ -108,6 +108,7 @@ meyer_wst(const Image& input_, const Neighborhood<N>& nbh_, L& nbasins) { + trace::entering("morpho::meyer_wst"); /* FIXME: Ensure the input image has scalar values. */ const I input = exact(input_); @@ -192,6 +193,7 @@ } } } + trace::exiting("morpho::meyer_wst"); return output; }

1 0

cleanup-2008 2671: Add a processing chain about segmentation on edges.
by Thierry Geraud 24 Oct '08

24 Oct '08

https://svn.lrde.epita.fr/svn/oln/branches/cleanup-2008/milena/sandbox Index: ChangeLog from Thierry Geraud <thierry.geraud(a)lrde.epita.fr> Add a processing chain about segmentation on edges. * geraud/wst_edge.cc: New. wst_edge.cc | 406 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 406 insertions(+) Index: geraud/wst_edge.cc --- geraud/wst_edge.cc (revision 0) +++ geraud/wst_edge.cc (revision 0) @@ -0,0 +1,406 @@ +# include <mln/core/var.hh> + +# include <mln/core/image/image2d.hh> +# include <mln/core/image/image_if.hh> +# include <mln/core/image/extended.hh> +# include <mln/core/routine/extend.hh> + +# include <mln/core/alias/window2d.hh> +# include <mln/core/alias/neighb2d.hh> +# include <mln/make/double_neighb2d.hh> +# include <mln/core/site_set/p_centered.hh> + +# include <mln/literal/origin.hh> +# include <mln/literal/black.hh> +# include <mln/literal/white.hh> + +# include <mln/value/int_u8.hh> +# include <mln/io/pgm/load.hh> +# include <mln/io/pgm/save.hh> + +# include <mln/value/rgb8.hh> +# include <mln/io/ppm/save.hh> + +# include <mln/accu/min_max.hh> +# include <mln/accu/mean.hh> + +# include <mln/fun/i2v/array.hh> +# include <mln/fun/p2v/iota.hh> + +# include <mln/level/paste.hh> +# include <mln/level/fill.hh> +# include <mln/level/transform.hh> +# include <mln/extension/fill.hh> + +# include <mln/morpho/meyer_wst.hh> +# include <mln/morpho/closing_area.hh> + +# include <mln/debug/println.hh> + + + +// Functions + +inline +bool is_row_odd(const mln::point2d& p) +{ + return p.row() % 2; +} + +inline +bool is_cell(const mln::point2d& p) +{ + return p.row() % 2 == 0 && p.col() % 2 == 0; +} + +inline +bool is_edge(const mln::point2d& p) +{ + return p.row() % 2 + p.col() % 2 == 1; +} + +inline +bool is_point(const mln::point2d& p) +{ + return p.row() % 2 && p.col() % 2; +} + + +inline +bool is_not_edge(const mln::point2d& p) +{ + return ! is_edge(p); +} + + + +namespace mln +{ + + namespace border + { + + template <typename I> + void + fill(I& ima, const mln_value(I)& v) + { + const int nrows = ima.nrows(); + const int ncols = ima.ncols(); + for (int r = -1; r <= nrows; ++r) + { + ima.at(r, -1) = v; + ima.at(r, ncols) = v; + } + for (int c = -1; c <= ncols; ++c) + { + ima.at(-1, c) = v; + ima.at(nrows, c) = v; + } + } + + } // mln::border + + namespace accu + { + + template <typename I, typename L, typename A, typename V> + inline + void + compute(const Image& input_, + const Image<L>& label_, + const Accumulator<A>&, + V& v) + { + trace::entering("accu::compute"); + + const I& input = exact(input_); + const L& label = exact(label_); + + const unsigned n = v.size(); + std::vector<A> a(n); + + mln_piter(I) p(input.domain()); + for_all(p) + a[label(p)].take(input(p)); + + for (unsigned l = 1; l < n; ++l) + v(l) = a[l].to_result(); + + trace::exiting("accu::compute"); + } + + } // mln::accu + + namespace morpho + { + + template <typename I, typename N> + mln_concrete(I) + gradient(const I& input, const N& nbh) + { + mln_concrete(I) output; + initialize(output, input); + accu::min_max<mln_value(I)> mm; + + mln_piter(I) p(input.domain()); + mln_niter(N) n(nbh, p); + for_all(p) + { + mm.init(); + for_all(n) if (input.has(n)) + mm.take(input(n)); + output(p) = mm.second() - mm.first(); + } + return output; + } + + template <typename I, typename N> + mln_concrete(I) + dilation(const I& input, const N& nbh) + { + typedef mln_value(I) V; + // FIXME: extension::fill(input, mln_min(V)); + + mln_concrete(I) output; + initialize(output, input); + accu::max<V> m; + + mln_piter(I) p(input.domain()); + mln_niter(N) n(nbh, p); + for_all(p) + { + m.init(); + for_all(n) if (input.has(n)) + m.take(input(n)); + output(p) = m; + } + return output; + } + + } // mln::morpho + + + struct colorize : Function_v2v< colorize > + { + typedef value::rgb8 result; + colorize(unsigned max) + : lut(max + 1) + { + lut[0] = literal::black; + for (unsigned i = 1; i <= max; ++i) + lut[i] = result(100 + std::rand() % 150, + 100 + std::rand() % 150, + 100 + std::rand() % 150); + } + result operator()(unsigned i) const + { + return lut[i]; + } + std::vector<result> lut; + }; + + + template <typename I> + I display_edge(const I& ima, mln_value(I) bg, unsigned zoom) + { + unsigned nrows = ima.nrows() / 2 + 1; + unsigned ncols = ima.ncols() / 2 + 1; + I output(nrows * (zoom + 1) - 1, + ncols * (zoom + 1) - 1); + level::fill(output, bg); + mln_VAR( edge, ima | is_edge ); + mln_piter(edge_t) p(edge.domain()); + for_all(p) + if (p.row() % 2) // horizontal edge + { + unsigned row = (p.row() / 2 + 1) * (zoom + 1) - 1; + unsigned col = (p.col() / 2) * (zoom + 1); + for (unsigned i = 0; i < zoom; ++i) + output.at(row, col + i) = ima(p); + } + else // vertical edge + { + unsigned row = (p.row() / 2) * (zoom + 1); + unsigned col = (p.col() / 2 + 1) * (zoom + 1) - 1; + for (unsigned i = 0; i < zoom; ++i) + output.at(row + i, col) = ima(p); + } + return output; + } + +} // mln + + + +template <typename T> +mln::image2d<T> +image2cells(const mln::image2d<T>& input) +{ + mln::image2d<T> output(2 * input.nrows() - 1, + 2 * input.ncols() - 1); + for (int row = 0; row < input.nrows(); ++row) + for (int col = 0; col < input.ncols(); ++col) + output.at(2 * row, 2 * col) = input.at(row, col); + return output; +} + + +template <typename T> +mln::image2d<T> +cells2image(const mln::image2d<T>& input) +{ + mln::image2d<T> output((input.nrows() + 1) / 2, + (input.ncols() + 1) / 2); + for (int row = 0; row < input.nrows(); row += 2) + for (int col = 0; col < input.ncols(); col += 2) + output.at(row / 2, col / 2) = input.at(row, col); + return output; +} + + + + +template <typename I> +void +do_it(I& ima, int lambda, const std::string& filename, unsigned& nbasins) +{ + using namespace mln; + + // e2c + + bool e2c_h[] = { 0, 1, 0, + 0, 0, 0, + 0, 1, 0 }; + + bool e2c_v[] = { 0, 0, 0, + 1, 0, 1, + 0, 0, 0 }; + + mln_VAR( e2c, make::double_neighb2d(is_row_odd, e2c_h, e2c_v) ); + + bool e2e_h[] = { 0, 0, 1, 0, 0, + 0, 1, 0, 1, 0, + 0, 0, 0, 0, 0, + 0, 1, 0, 1, 0, + 0, 0, 1, 0, 0 }; + + bool e2e_v[] = { 0, 0, 0, 0, 0, + 0, 1, 0, 1, 0, + 1, 0, 0, 0, 1, + 0, 1, 0, 1, 0, + 0, 0, 0, 0, 0 }; + + mln_VAR( e2e, make::double_neighb2d(is_row_odd, e2e_h, e2e_v) ); + + // cell + mln_VAR(cell, ima | is_cell); + + // edge + mln_VAR(edge, extend((ima | is_edge).rw(), + pw::value(ima))); + + level::paste(morpho::gradient(edge, e2c), edge); + // ^^^ + // edge -> neighboring cells + + // 'edge' looks like: + // + // 1 1 + // 3 3 3 + // 1 1 + + { + io::pgm::save(display_edge(ima, 0, 3), + "temp_edge.pgm"); + } + + + level::paste( morpho::closing_area(edge, e2e, lambda), edge ); + + + image2d<unsigned> label(ima.bbox(), 0); + // mln_ch_value(I, unsigned) label; + // initialize(label, ima); + + + mln_VAR(wst, label | is_edge); + + level::fill(wst, morpho::meyer_wst(edge, e2e, nbasins)); + // ^^^ + // edge -> neighboring edges + + // 'wst' looks like: + // + // 2 2 + // 0 0 0 + // 1 1 + + colorize colors(nbasins); + + { + image2d<value::rgb8> temp(label.domain()); + level::fill(temp, literal::white); + + level::paste( level::transform(label | is_edge, + colors), + temp ); + + io::ppm::save(display_edge(temp, literal::white, 3), + "temp_label.ppm"); + } + + // YET THOSE VALUES ARE ON EDGES, NOT ON CELLS... + + + mln_VAR(lab, label | is_cell); + + level::paste(morpho::dilation(extend(lab, label), + c4()), + label); + + // 'lab' looks like: + // + // 2 2 2 + // + // 1 1 1 + + io::ppm::save(level::transform(cells2image(label), + colors), + filename); +} + + + +void usage(char* argv[]) +{ + std::cerr << "usage: " << argv[0] << " input.pgm lambda output.ppm" << std::endl; + std::cerr << " lambda >= 0" << std::endl; + abort(); +} + + + +int main(int argc, char* argv[]) +{ + using namespace mln; + using value::int_u8; + + if (argc != 4) + usage(argv); + + image2d<int_u8> temp; + io::pgm::load(temp, argv[1]); + + image2d<int_u8> ima = image2cells(temp); + + + int lambda = atoi(argv[2]); + if (lambda < 0) + usage(argv); + + std::string filename(argv[3]); + + unsigned nbasins; + do_it(ima, lambda, filename, nbasins); + std::cout << "nbasins = " << nbasins << std::endl; +}

1 0

r2670: Fix the skeleton algorithm introduced by aroumougame
by Jimmy Ma 24 Oct '08

24 Oct '08

URL: https://svn.lrde.epita.fr/svn/oln/branches/cleanup-2008/milena/sandbox ChangeLog: 2008-10-24 Jimmy Ma <jimmy.ma(a)lrde.epita.fr> Fix the skeleton algorithm introduced by aroumougame. * garrigues/ocr/skeleton.cc, garrigues/ocr/skeleton.hh: New. The code has been updated to work with the current version of milena. However, the code itself has to be cleaned and improved. --- skeleton.cc | 30 ++ skeleton.hh | 611 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 641 insertions(+) Index: branches/cleanup-2008/milena/sandbox/garrigues/ocr/skeleton.hh =================================================================== --- branches/cleanup-2008/milena/sandbox/garrigues/ocr/skeleton.hh (revision 0) +++ branches/cleanup-2008/milena/sandbox/garrigues/ocr/skeleton.hh (revision 2670) @@ -0,0 +1,611 @@ +// Copyright (C) 2007, 2008 EPITA Research and Development Laboratory +// +// This file is part of the Olena Library. This library is free +// software; you can redistribute it and/or modify it under the terms +// of the GNU General Public License version 2 as published by the +// Free Software Foundation. +// +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +// General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this library; see the file COPYING. If not, write to +// the Free Software Foundation, 51 Franklin Street, Fifth Floor, +// Boston, MA 02111-1307, USA. +// +// As a special exception, you may use this file as part of a free +// software library without restriction. Specifically, if other files +// instantiate templates or use macros or inline functions from this +// file, or you compile this file and link it with other files to +// produce an executable, this file does not by itself cause the +// resulting executable to be covered by the GNU General Public +// License. This exception does not however invalidate any other +// reasons why the executable file might be covered by the GNU General +// Public License. +#ifndef SKELETON_HH +# define SKELETON_HH + +#include <mln/core/image/image2d.hh> +#include <mln/core/alias/neighb2d.hh> +#include <sandbox/aroumougame/skeleton/sedt.hh> + +#include <mln/core/site_set/p_set.hh> +#include <mln/math/sqrt.hh> + +namespace mln +{ + + +template <typename P> + std::vector< std::pair< double, P > > remove(std::vector< std::pair< double, P > > Q, P p) +{ + typename std::vector<std::pair< double, P > >::iterator it; + + for(it = Q.begin(); it!=Q.end(); it++) + { + if((*it).second==p) + { + it = Q.erase(it); + break; + } + } + return Q; +} +template <typename N> + double distance(N a, N b, N c, N d) +{ + double dist = sqrt((a-c)*(a-c)+(b-d)*(b-d)); + return dist; +} +template <typename P> + std::vector< std::pair< double, P > > insertDicho(std::vector< std::pair< double, P > > Q, std::pair< double, P> p) +{ + int indMin, indMax, indMid; + + indMin = 0; + indMax = Q.size(); + + if(indMax==0 || Q[indMax-1].first <= p.first) + Q.push_back(p); + else + { + while(indMax > indMin) + { + indMid = int(indMin + (indMax-indMin)/2); + if(Q[indMid].first < p.first) + { + indMin = indMid+1; + if(Q[indMin].first > p.first) + { + indMax = indMid; + } + } + else + { + indMax = indMid-1; + if(Q[indMax].first < p.first) + { + indMin = indMid; + } + } + } + + typename std::vector< std::pair< double, P > >::iterator it=Q.begin(); + it = Q.insert ( it+indMin, p); + } + + return Q; +} + + +const neighb2d& complement_neighborhood(const Neighborhood<neighb2d>& nbh) +{ + if(&nbh == &c4()) + return c8(); + return c4(); +} + +template <typename N> + int nb_composant_connexe(p_set< mln_psite(N) > X_full,const Neighborhood<N>& nbh_,const mln_psite(N)& p_ref, bool local) +{ + N nbh = exact(nbh_); + p_set< mln_psite(N) > X; + if(local) + { + mln_niter(N) n(max_neighborhood(nbh), p_ref); + + for_all(n) + { + if (X_full.has(n)) + X.insert(n); + } + } + else + { + X = X_full; + } + + int T; + p_set< mln_psite(N) > done; + p_set< mln_psite(N) > neighbors; + p_set< mln_psite(N) > composant; + + done.insert(p_ref); + mln_niter(N) q(nbh, p_ref); + for_all(q) + { + if (X.has(q)&&!done.has(q)) + { + neighbors.insert(q); + } + } +// std::cout << "nb_composant_connexe: neighbors " << neighbors.nsites() <<std::endl; + if(neighbors.nsites()<=1) + { + return neighbors.nsites(); + } + else + T=0; + + while(neighbors.nsites()!=0) + { + T++; + done.insert(neighbors[0]); + mln_niter(N) t(nbh, neighbors[0]); + neighbors.remove(neighbors[0]); + for_all(t) + { + if (X.has(t)&&!done.has(t)) + { + composant.insert(t); + } + } + + while(composant.nsites()!=0) + { + done.insert(composant[0]); + if(neighbors.has(composant[0])) + { + neighbors.remove(composant[0]); + if(neighbors.nsites()==0) + return T; + } + + mln_niter(N) r(nbh, composant[0]); + composant.remove(composant[0]); + for_all(r) + { + if (X.has(r) && !done.has(r)) + { + composant.insert(r); + } + } + } + } + return T; +} + +template <typename N> + bool simple_point(p_set< mln_psite(N) > X,const Neighborhood<N>& nbh, p_set< mln_psite(N) > X_complement, const mln_psite(N)& p_ref, bool local) +{ + int nX = nb_composant_connexe(X,exact(nbh),p_ref,local); + int nX_complement = nb_composant_connexe(X_complement,complement_neighborhood(exact(nbh)),p_ref,local); + + if((nX_complement == 1)&&(nX == 1)) + return true; + return false; +} + + + template <typename N> + p_set<mln_psite(N)> euclideanSkeleton(p_set<mln_psite(N)> X, p_set<mln_psite(N)> X_complement, image2d<value::int_u8> dt, p_set<mln_psite(N)>& Y, const Neighborhood<N>& nbh_, bool local) + { + std::vector< std::pair< double, mln::point2d> > Q; + std::vector< std::pair< double, mln::point2d> > R; + N nbh = exact(nbh_); + + // fill Q + for (uint i = 0; i < X.nsites(); i++) + { + if (!Y.has(X[i])) + { + std::pair<double, mln_psite(N)> p(math::sqrt(double(dt(X[i]))),X[i]); + Q = insertDicho(Q,p); + } + } + + // fill R + for (uint i = 0; i < X.nsites(); i++) + { + if (!Y.has(X[i])) + { + double min = 1023.99; + mln_niter(N) r(nbh, X[i]); + for_all(r) + { + if (Y.has(r)) + { + double tmp = distance(r[0], r[1], X[i][0], X[i][1]); + double d = math::sqrt(double(dt(r)))+(math::sqrt(double(dt(X[i])))-math::sqrt(double(dt(r))))/tmp; + min = math::min(min,d); + } + } + if (min!=1023.99) + { + std::pair<double, mln_psite(N)> p(min,X[i]); + R = insertDicho(R, p); + } + } + } + + while (!Q.empty() || !R.empty()) + { + mln_psite(N) tmp; + if (Q[0].first < R[0].first) + { + tmp = Q[0].second; + } + else + { + tmp = R[0].second; + } + + Q = remove(Q, tmp); + R = remove(R, tmp); + + if (!Y.has(tmp) && X.has(tmp)) + { + if (simple_point(X, nbh, X_complement, tmp, local)) + { + X.remove(tmp); + X_complement.insert(tmp); + } + else + { + Y.insert(tmp); + mln_niter(N) r(nbh, tmp); + for_all(r) + { + if (!Y.has(r) && X.has(r)) + { + double dist = distance(r[0], r[1], tmp[0], tmp[1]); + double d = math::sqrt(double(dt(tmp)))+(math::sqrt(double(dt(r)))-math::sqrt(double(dt(tmp))))/dist; + std::pair<double, mln_psite(N)> p(d,r); + R = insertDicho(R, p); + } + } + + } + } + + } + return X; + } + + + p_set<point2d> EP(image2d<value::int_u8> dt, point2d x, std::vector< std::vector<std::pair< int, int> > > lut, const neighb2d& nbh) + { + p_set<point2d> EP; + p_set<point2d> tmp; + int w = geom::ncols(dt); + int h = geom::nrows(dt); + + + mln_niter_(neighb2d) r(nbh, x); + for_all(r) + { + if (dt(r) <= dt(x)) + { + for (uint i=0; i<lut[dt(r)].size(); i++) + { + if ((r[0]+lut[dt(r)][i].first < h) && (r[1]+lut[dt(r)][i].second < w)) + { + if (!dt(r+dpoint2d(lut[dt(r)][i].first, lut[dt(r)][i].second))) + EP.insert(r+dpoint2d(lut[dt(r)][i].first, lut[dt(r)][i].second)); + } + if ((r[0]-lut[dt(r)][i].first >= 0) && (r[1]-lut[dt(r)][i].second >= 0)) + { + if (!dt(r+dpoint2d(-lut[dt(r)][i].first, -lut[dt(r)][i].second))) + EP.insert(r+dpoint2d(-lut[dt(r)][i].first, -lut[dt(r)][i].second)); + } + if ((r[0]+lut[dt(r)][i].first < h) && (r[1]-lut[dt(r)][i].second >= 0)) + { + if (!dt(r+dpoint2d(lut[dt(r)][i].first, -lut[dt(r)][i].second))) + EP.insert(r+dpoint2d(lut[dt(r)][i].first, -lut[dt(r)][i].second)); + } + if ((r[0]-lut[dt(r)][i].first >= 0) && (r[1]+lut[dt(r)][i].second < w)) + { + if (!dt(r+dpoint2d(-lut[dt(r)][i].first, lut[dt(r)][i].second))) + EP.insert(r+dpoint2d(-lut[dt(r)][i].first, lut[dt(r)][i].second)); + } + if ((r[0]+lut[dt(r)][i].second < h) && (r[1]+lut[dt(r)][i].first < w)) + { + if (!dt(r+dpoint2d(lut[dt(r)][i].second, lut[dt(r)][i].first))) + EP.insert(r+dpoint2d(lut[dt(r)][i].second, lut[dt(r)][i].first)); + } + if ((r[0]-lut[dt(r)][i].second >= 0) && (r[1]-lut[dt(r)][i].first >= 0)) + { + if (!dt(r+dpoint2d(-lut[dt(r)][i].second, -lut[dt(r)][i].first))) + EP.insert(r+dpoint2d(-lut[dt(r)][i].second, -lut[dt(r)][i].first)); + } + if ((r[0]+lut[dt(r)][i].second < h) && (r[1]-lut[dt(r)][i].first >= 0)) + { + if (!dt(r+dpoint2d(lut[dt(r)][i].second, -lut[dt(r)][i].first))) + EP.insert(r+dpoint2d(lut[dt(r)][i].second, -lut[dt(r)][i].first)); + } + if ((r[0]-lut[dt(r)][i].second >= 0) && (r[1]+lut[dt(r)][i].first < w)) + { + if (!dt(r+dpoint2d(-lut[dt(r)][i].second, lut[dt(r)][i].first))) + EP.insert(r+dpoint2d(-lut[dt(r)][i].second, lut[dt(r)][i].first)); + } + } + } + } + + return EP; + } + + std::vector< std::vector<std::pair< int, int> > > Lut2d(int N) + { + int n = int(sqrt(N))+1; + int i=0; + std::vector< std::vector<std::pair< int, int> > > lut; + + for(i = 0; i <= N; i++) + { + std::vector<std::pair< int, int> > vect; + lut.push_back(vect); + } + + for(int x = 0; x <= n; x++) + { + for(int y = 0; y <= x; y++) + { + i=x*x+y*y; + if(i<=N) + { + std::pair<int,int> p(x,y); + lut[i].push_back(p); + } + } + } + + return lut; +} + + image2d<value::int_u8> DiscreteBisector(image2d<value::int_u8> dt, p_set<point2d> Y, const neighb2d& nbh, int N) + { + int w = geom::ncols(dt); + int h = geom::nrows(dt); + + int ux,uy,vx,vy, produit, angle, max; + double cos, normu, normv; + + std::vector< std::vector<std::pair< int, int> > > lut; + lut = Lut2d(N); + + p_set<point2d> proj; + + image2d<value::int_u8> bisector(h, w); + level::fill(bisector, 0); + + for (uint i=0; i<Y.nsites(); i++) + { + proj = EP(dt, Y[i], lut, nbh); + + int n=proj.nsites(); + + if (n>1) + { + max = 0; + for (int y=0; y<n; y++) + { + for (int z=0; z<y; z++) + { + ux = proj[y][0]-Y[i][0]; + uy = proj[y][1]-Y[i][1]; + vx = proj[z][0]-Y[i][0]; + vy = proj[z][1]-Y[i][1]; + + produit = ux * vx + uy * vy; + + normu = sqrt(ux*ux + uy*uy); + normv = sqrt(vx*vx + vy*vy); + + cos = produit/(normu*normv); + angle = int(acos(cos)*180.0/3.1415); + + max = math::max(max, angle); + } + } + bisector(Y[i]) = max; + } + + } + + return bisector; + + } + + + +const neighb2d& max_neighborhood(const Neighborhood<neighb2d>& nbh) +{ + return c8(); +} +template <typename N> + p_set<mln_psite(N)> ultimateSkeleton(p_set<mln_psite(N)> X, p_set<mln_psite(N)> X_complement, image2d<value::int_u8> dt, p_set<mln_psite(N)> Y, const Neighborhood<N>& nbh_, bool local) +{ + std::vector< std::pair< double, mln::point2d> > Q; + + N nbh = exact(nbh_); + // fill Q + for(uint i = 0; i < X.nsites(); i++) + { + if (!Y.has(X[i])) + { + std::pair<double, mln_psite(N)> p(dt(X[i]),X[i]); + Q = insertDicho(Q,p); + } + } + + + while(!Q.empty()) + { + mln_psite(N) tmp = Q[0].second; + + Q = remove(Q, tmp); + + if(simple_point(X, nbh, X_complement, tmp, local)) + { + X.remove(tmp); + X_complement.insert(tmp); + mln_niter(N) r(nbh, tmp); + for_all(r) + { + if(!Y.has(r) && X.has(r)) + { + std::pair<double, mln_psite(N)> p(dt(r),r); + Q = insertDicho(Q, p); + } + } + } + + } + return X; +} + + image2d<value::int_u8> intImage(image2d<bool> pic) +{ + int w = geom::ncols(pic); + int h = geom::nrows(pic); + + image2d<value::int_u8> out(h,w); + for(int i=0; i<w; i++) + for(int j=0; j<h; j++) + { + if(pic.at(j,i)) + out.at(j,i) = 1; + else + out.at(j,i) = 0; + } + return out; +} + image2d<bool> filteredSkeleton(image2d<bool> pic, const neighb2d& nbh, uint r, uint alpha, bool local) + { + using value::int_u8; + + typedef image2d<bool> I; + typedef p_set<point2d> S; + + image2d<value::int_u8> pic1 = intImage(pic); + image2d<value::int_u8> dt = sedt(pic1); + + mln::io::pgm::save(dt, "dt.pgm"); + + int w = geom::ncols(pic); + int h = geom::nrows(pic); + int l= math::min(w, h); + uint rmax = getRMax(dt); + uint rknown = 0; + p_set<point2d> X,Y,Z; + p_set<point2d> X_complement, Z_complement; + + image2d<value::int_u8> DTg(l,l,0); + std::vector< std::vector<int> > Mgl; + std::vector< std::vector<int> > Lut; + + Mgl = CompLutMask (DTg,Mgl,Lut,l,0,rmax); + + rknown =rmax; + + mln_fwd_piter_(image2d<bool>) p(pic.domain()); + + for_all(p) + { + if (pic(p)==1) + { + X.insert(p); + } + else + { + X_complement.insert(p); + } + } + std::cout << " medial axis " << std::endl; + pic = MA(pic, Mgl, dt, Lut); + + mln::io::pbm::save(pic, "ma.pbm"); + + for_all(p) + { + if (pic(p)==1) + { + Y.insert(p); + } + } + + std::cout << " euclidean skeleton " << std::endl; + Z = euclideanSkeleton(X, X_complement, dt, Y, nbh, local); + + sub_image<I, S> es = pic | Z; + I es1(pic.domain()); + level::fill(es1, false); + + level::paste(es, es1); + + mln::io::pbm::save(es1, "euclidean.pbm"); + + for_all(p) + { + if (!Z.has(p)) + { + Z_complement.insert(p); + } + } + std::cout << " discrete bisector " << std::endl; + pic1 = DiscreteBisector(dt, Y, nbh, rmax); + + + mln::io::pgm::save(pic1, "bisector.pgm"); + + uint cpt=0; + while (cpt!=Y.nsites()) + { + if (dt(Y[cpt])>=r && pic1(Y[cpt])>=alpha) + { + cpt++; + } + else + { + Y.remove(Y[cpt]); + } + } + + + sub_image<I, S> skel = pic | Y; + I test(pic.domain()); + level::fill(test, false); + + level::paste(skel, test); + + mln::io::pbm::save(test, "Y.pbm"); + + std::cout << " ultimate skeleton " << std::endl; + Z = ultimateSkeleton(Z, Z_complement, dt, Y, nbh, local); + + + + sub_image<I, S> skeleton = pic | Z; + I output(pic.domain()); + level::fill(output, false); + + level::paste(skeleton, output); + + return output; + } + +} // End of namespace mln +#endif // ! SKELETON_HH Index: branches/cleanup-2008/milena/sandbox/garrigues/ocr/skeleton.cc =================================================================== --- branches/cleanup-2008/milena/sandbox/garrigues/ocr/skeleton.cc (revision 0) +++ branches/cleanup-2008/milena/sandbox/garrigues/ocr/skeleton.cc (revision 2670) @@ -0,0 +1,30 @@ + +#include <mln/core/alias/point2d.hh> +#include "skeleton.hh" +#include <mln/level/paste.hh> +#include <mln/level/fill.hh> +#include <mln/core/image/sub_image.hh> +#include <mln/io/pgm/save.hh> +#include <mln/io/pbm/save.hh> +#include <mln/io/pbm/load.hh> + + +int main(int argc, char* argv[]) +{ + if(argc!=5) + { + std::cout << "arguments: filename voisinage R alpha" << std::endl; + exit(1); + } + image2d<bool> output; + std::string filename = argv[1]; + int r = atoi(argv[3]); + int alpha = atoi(argv[4]); + + image2d<bool> pic = io::pbm::load(filename); + if(atoi(argv[2])==4) + output = filteredSkeleton( pic, c4(), r, alpha, true); + else + output = filteredSkeleton( pic, c8(), r, alpha, true); + mln::io::pbm::save(output, "FS-"+std::string(argv[2])+"_"+std::string(argv[3])+"_"+std::string(argv[4])+"_"+filename); +}

1 0

r2669: Update OCR using tesseract
by Jimmy Ma 24 Oct '08

24 Oct '08

URL: https://svn.lrde.epita.fr/svn/oln/branches/cleanup-2008/milena/sandbox ChangeLog: 2008-10-24 Jimmy Ma <jimmy.ma(a)lrde.epita.fr> Update OCR using tesseract. * garrigues/ocr/enlarge.hh: New. New enlargement algorithm better than the naive resize. * garrigues/ocr/ocr.cc: Provide sample code to link milena against tesseract. --- enlarge.hh | 121 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ocr.cc | 91 ++++++++++++++++++++++++++++++++++++++------- 2 files changed, 198 insertions(+), 14 deletions(-) Index: branches/cleanup-2008/milena/sandbox/garrigues/ocr/ocr.cc =================================================================== --- branches/cleanup-2008/milena/sandbox/garrigues/ocr/ocr.cc (revision 2668) +++ branches/cleanup-2008/milena/sandbox/garrigues/ocr/ocr.cc (revision 2669) @@ -36,40 +36,103 @@ #include <mln/value/int_u8.hh> #include "resize.hh" -#include <mln/linear/convolve.hh> +#include "enlarge.hh" +//#include "skeleton.hh" #include <mln/linear/gaussian.hh> #include <mln/trace/all.hh> -#include <mln/io/pbm/load.hh> +#include <mln/io/pgm/load.hh> #include <mln/io/pgm/save.hh> +#include <mln/io/pbm/load.hh> +#include <mln/io/pbm/save.hh> #include <mln/core/alias/w_window2d_float.hh> +#include <mln/debug/println.hh> +#include <mln/geom/chamfer.hh> +#include <mln/make/win_chamfer.hh> +#include <mln/labeling/regional_maxima.hh> +#include <mln/morpho/dilation.hh> + +#include <tesseract/baseapi.h> + +// _COMPILATION_ +// g++ -DNDEBUG -O3 -I../../.. ocr.cc -L/usr/lib -ltesseract_full -lpthread + + +// Call tesseract +// lang: expected language +template <typename T> +char* tesseract(const char* lang, const mln::image2d<T>& input) +{ + TessBaseAPI::InitWithLanguage(NULL, NULL, lang, NULL, false, 0, NULL); + char* s = TessBaseAPI::TesseractRect( + (unsigned char*) input.buffer(), + sizeof (T), + input.ncols() * sizeof (T), + 0, 0, + input.ncols(), + input.nrows()); + return s; +} int main(int argc, char** argv) { - mln::trace::quiet = false; - using namespace mln; using value::int_u8; image2d<bool> input; - image2d<int_u8> output; - image2d<int_u8> output2; - if (argc != 3) + if (argc < 2) return 1; - + mln::border::thickness = 0; io::pbm::load(input, argv[1]); + std::cout << "> without any preprocessing." << std::endl; + char* s = tesseract("fra", input); + std::cout << s; + free(s); + + // + // PREPROCESS !! + // + // Resize +// output2 = geom::resize(cast_image<int_u8>(input), 3); + image2d<int_u8> output = enlarge(input, 1); + + // TODO CLEANUP +#if 0 + // Blur. + image2d<int_u8> output; + initialize(output, output2); + linear::gaussian(output2, 1, output); +#endif + +#if 0 + // Threshold + mln_piter_(image2d<unsigned>) p(output.domain()); + for_all(p) + { + output(p) = output(p) > 127 ? 1 : 0; + } +#endif - // Step 1: Enlarge input. - output = geom::resize(cast_image<int_u8>(input), 4); +#if 0 + // Compute chamfer distance map. + const w_window2d_int& w_win = make::mk_chamfer_3x3_int<8, 0> (); + image2d<unsigned> out = geom::chamfer(output, w_win, 255); + + for_all(p) + { + out(p) = out(p) > 10 ? 255 : 0; + } +#endif - // Step 2: Blur. - initialize(output2, output); - linear::gaussian(output, 15, output2); + io::pgm::save(cast_image<int_u8>(output), argv[2]); - io::pgm::save(output2, argv[2]); + std::cout << "> with preprocessing." << std::endl; + s = tesseract("fra", output); + std::cout << s; + free(s); } Index: branches/cleanup-2008/milena/sandbox/garrigues/ocr/enlarge.hh =================================================================== --- branches/cleanup-2008/milena/sandbox/garrigues/ocr/enlarge.hh (revision 0) +++ branches/cleanup-2008/milena/sandbox/garrigues/ocr/enlarge.hh (revision 2669) @@ -0,0 +1,121 @@ +#ifndef ENLARGE_HH +# define ENLARGE_HH + +# include <iostream> + +# include <mln/core/image/image2d.hh> +# include <mln/value/int_u8.hh> + + +float val(bool b) { return b ? 1 : 0; } + +int do_threshold(float value) +{ + return 255.f * value; +} + + +mln::image2d<mln::value::int_u8> +enlargex2(mln::image2d<bool> input) +{ + using namespace mln; + using value::int_u8; + + unsigned nrows, ncols; + + nrows = input.nrows(); + ncols = input.ncols(); + + image2d<int_u8> output(2 * nrows, 2 * ncols); + float value; + + // row 0 + + output.at(0, 0) = do_threshold(input.at(0, 0)); + + for (int col = 2; col < output.ncols(); col += 2) + { + value = val(input.at(0, col / 2)); + value += val(input.at(0, col / 2 - 1)); + output.at(0, col) = do_threshold(value / 2); + } + + for (int col = 1; col < output.ncols(); col += 2) + output.at(0, col) = do_threshold(input.at(0, col / 2)); + + // col 0 + + for (int row = 2; row < output.nrows(); row += 2) + { + value = val(input.at(row / 2, 0)); + value += val(input.at(row / 2 - 1, 0)); + output.at(row, 0) = do_threshold(value / 2); + } + + for (int row = 1; row < output.nrows(); row += 2) + output.at(row, 0) = do_threshold(input.at(row / 2, 0)); + + // others + + for (int row = 2; row < output.nrows(); row += 2) + { + for (int col = 2; col < output.ncols(); col += 2) + { + value = val(input.at(row / 2, col / 2)); + value += val(input.at(row / 2 - 1, col / 2)); + value += val(input.at(row / 2, col / 2 - 1)); + value += val(input.at(row / 2 - 1, col / 2 - 1)); + output.at(row, col) = do_threshold(value / 4); + } + for (int col = 1; col < output.ncols(); col += 2) + { + value = val(input.at(row / 2, col / 2)); + value += val(input.at(row / 2 - 1, col / 2)); + output.at(row, col) = do_threshold(value / 2); + } + } + + for (int row = 1; row < output.nrows(); row += 2) + { + for (int col = 2; col < output.ncols(); col += 2) + { + value = val(input.at(row / 2, col / 2)); + value += val(input.at(row / 2, col / 2 - 1)); + output.at(row, col) = do_threshold(value / 2); + } + for (int col = 1; col < output.ncols(); col += 2) + output.at(row, col) = do_threshold(input.at(row / 2, col / 2)); + } + + return output; +} + + +// enlarge 2^n times +mln::image2d<mln::value::int_u8> +enlarge(mln::image2d<bool> input, unsigned int n) +{ + using namespace mln; + using value::int_u8; + + image2d<int_u8> output; + + do + { + output = enlargex2(input); + + if (--n > 0) + { + initialize(input, output); + mln_piter_(image2d<bool>) p(input.domain()); + for_all(p) + input(p) = output(p) > 0; + } + else + break; + } + while (1); + return output; +} + +#endif // ! ENLARGE_HH

1 0

cleanup-2008 2668: Update directional erosion so that it can now be used on sets.
by Thierry Geraud 23 Oct '08

23 Oct '08

https://svn.lrde.epita.fr/svn/oln/branches/cleanup-2008/milena Index: ChangeLog from Thierry Geraud <thierry.geraud(a)lrde.epita.fr> Update directional erosion so that it can now be used on sets. * mln/accu/rank.hh (todo): New. * mln/accu/rank_bool.hh (todo): New. Fix doc. (rank): Fix ctor. (take_as_init): Fix. * mln/accu/land.hh: New. * mln/morpho/erosion.spe.hh (erosion_directional_nd_functor), (erosion_directional_nd_fastest_functor): Add accumulator type as parameter and rename min as accu. (erosion_directional_nd), (erosion_directional_nd_fastest): Update; compute A. (erosion_dispatch_wrt_win): Update overload for hline2d and vline2d. * mln/morpho/includes.hh: Update. accu/land.hh | 81 ++++++++++++++++++++++++-------------------------- accu/rank.hh | 3 + accu/rank_bool.hh | 18 +++++------ morpho/erosion.spe.hh | 58 ++++++++++++++++++----------------- morpho/includes.hh | 2 + 5 files changed, 83 insertions(+), 79 deletions(-) Index: mln/accu/rank.hh --- mln/accu/rank.hh (revision 2667) +++ mln/accu/rank.hh (working copy) @@ -31,6 +31,9 @@ /*! \file mln/accu/rank.hh * * \brief Define an rank accumulator. + * + * \todo It should be renamed as rank_h since it relies on histogram + * (thus low quantization). */ # include <vector> Index: mln/accu/rank_bool.hh --- mln/accu/rank_bool.hh (revision 2667) +++ mln/accu/rank_bool.hh (working copy) @@ -28,16 +28,17 @@ #ifndef MLN_ACCU_RANK_BOOL_HH # define MLN_ACCU_RANK_BOOL_HH -/*! \file mln/accu/rankbool.hh +/*! \file mln/accu/rank_bool.hh * * \brief Define an rank accumulator. + * + * \todo There is no-arg-ctor so this accumulator does not support + * deferred initialization! + * + * \todo Add untake routines... */ -# include <vector> # include <mln/accu/internal/base.hh> -# include <mln/core/concept/meta_accumulator.hh> -# include <mln/trait/value_.hh> -# include <mln/util/pix.hh> namespace mln @@ -49,7 +50,7 @@ // Fwd declaration. template <typename T> struct rank; - /*! \brief rank accumulator class for boolean. + /*! \brief rank accumulator class for Boolean. * */ template <> @@ -85,8 +86,7 @@ inline rank<bool>::rank(unsigned k, unsigned n) - : nfalse_(0), - k_(k), + : k_(k), n_(n) { mln_assertion(k_ < n_); @@ -105,7 +105,7 @@ inline void rank<bool>::take_as_init(const argument& t) { - nfalse_ += !t; + nfalse_ = t ? 0 : 1; } Index: mln/accu/land.hh --- mln/accu/land.hh (revision 2667) +++ mln/accu/land.hh (working copy) @@ -25,19 +25,15 @@ // reasons why the executable file might be covered by the GNU General // Public License. -#ifndef MLN_ACCU_RANK_BOOL_HH -# define MLN_ACCU_RANK_BOOL_HH +#ifndef MLN_ACCU_LAND_HH +# define MLN_ACCU_LAND_HH -/*! \file mln/accu/rankbool.hh +/*! \file mln/accu/land.hh * - * \brief Define an rank accumulator. + * \brief Define a 'logical-and' accumulator. */ -# include <vector> # include <mln/accu/internal/base.hh> -# include <mln/core/concept/meta_accumulator.hh> -# include <mln/trait/value_.hh> -# include <mln/util/pix.hh> namespace mln @@ -46,25 +42,23 @@ namespace accu { - // Fwd declaration. - template <typename T> struct rank; - - /*! \brief rank accumulator class for boolean. - * - */ - template <> - struct rank<bool> : public mln::accu::internal::base< bool, rank<bool> > + /// "Logical-and" accumulator class. + struct land : public mln::accu::internal::base< bool, land > { typedef bool argument; - rank(unsigned k, unsigned n); + land(); /// Manipulators. /// \{ void init(); void take_as_init(const argument& t); + void take(const argument& t); - void take(const rank<bool>& other); + void take(const land& other); + + void untake(const argument& t); + void untake(const land& other); /// \} /// Get the value of the accumulator. @@ -76,67 +70,71 @@ protected: unsigned nfalse_; - unsigned k_; // 0 <= k_ < n - unsigned n_; }; # ifndef MLN_INCLUDE_ONLY inline - rank<bool>::rank(unsigned k, unsigned n) - : nfalse_(0), - k_(k), - n_(n) + land::land() { - mln_assertion(k_ < n_); init(); } - inline void - rank<bool>::init() + land::init() { nfalse_ = 0; } - inline - void rank<bool>::take_as_init(const argument& t) + void land::take_as_init(const argument& t) { - nfalse_ += !t; + nfalse_ = t ? 0 : 1; } - inline - void rank<bool>::take(const argument& t) + void land::take(const argument& t) { - nfalse_ += !t; + if (t == false) + ++nfalse_; } - inline void - rank<bool>::take(const rank<bool>& other) + land::take(const land& other) { nfalse_ += other.nfalse_; } + inline + void land::untake(const argument& t) + { + if (t == false) + --nfalse_; + } + + inline + void + land::untake(const land& other) + { + mln_precondition(other.nfalse_ <= nfalse_); + nfalse_ -= other.nfalse_; + } inline bool - rank<bool>::to_result() const + land::to_result() const { - mln_assertion(nfalse_ <= n_); - return k_ >= nfalse_; + return nfalse_ == 0; } inline bool - rank<bool>::is_valid() const + land::is_valid() const { - return nfalse_ <= n_; + return true; } # endif // ! MLN_INCLUDE_ONLY @@ -145,5 +143,4 @@ } // end of namespace mln - -#endif // ! MLN_ACCU_RANK_BOOL_HH +#endif // ! MLN_ACCU_AND_HH Index: mln/morpho/erosion.spe.hh --- mln/morpho/erosion.spe.hh (revision 2667) +++ mln/morpho/erosion.spe.hh (working copy) @@ -636,7 +636,7 @@ return dp; } - template <typename I_, typename W> + template <typename I_, typename W, typename A> struct erosion_directional_nd_functor { typedef I_ I; @@ -645,7 +645,7 @@ const I& input; const W& win; mln_concrete(I) output; - accu::min_h<mln_value(I)> min; + A accu; enum { dim = I::site::dim }; mln_psite(I) p; @@ -662,7 +662,7 @@ erosion_directional_nd_functor(const I& input, const W& win, int dir) : input(input), win(win), - min(), + accu(), dir(dir), win_left(win::shift(win, -dp_directional<dpsite>(dir)) - win), win_right(win - win::shift(win, -dp_directional<dpsite>(dir))), @@ -675,27 +675,27 @@ void init() { // FIXME: border::adjust(input, win.delta()); - extension::fill(input, mln_max(mln_value(I))); + extension::fill(input, accu); initialize(output, input); } void init_run() { - min.init(); + accu.init(); p[dir]--; mln_qiter(W) q(win, p); for_all(q) if (input.has(q)) - min.take(input(q)); + accu.take(input(q)); p[dir]++; } void next() { for_all(q_l) if (input.has(q_l)) - min.untake(input(q_l)); + accu.untake(input(q_l)); for_all(q_r) if (input.has(q_r)) - min.take(input(q_r)); - output(p) = min; + accu.take(input(q_r)); + output(p) = accu; } void final() @@ -711,7 +711,11 @@ { trace::entering("morpho::impl:erosion_directional_nd"); - typedef erosion_directional_nd_functor<I, W> F; + typedef mlc_is(mln_trait_image_kind(I), + trait::image::kind::binary) is_binary; + typedef mlc_if(is_binary, accu::land, accu::min_h<mln_value(I)>) A; + + typedef erosion_directional_nd_functor<I, W, A> F; F f(exact(input), exact(win), dir); canvas::browsing::directional(f); @@ -721,7 +725,7 @@ } - template <typename I_, typename W> + template <typename I_, typename W, typename A> struct erosion_directional_nd_fastest_functor { typedef I_ I; @@ -730,7 +734,7 @@ const I& input; const W& win; mln_concrete(I) output; - accu::min_h<mln_value(I)> min; + A accu; mln_psite(I) p; enum { dim = I::site::dim }; @@ -743,7 +747,7 @@ erosion_directional_nd_fastest_functor(const I& input, const W& win, unsigned dir) : input(input), win(win), - min(), + accu(), dir(dir), win_left(win::shift(win, -dp_directional<dpsite>(dir)) - win), win_right(win - win::shift(win, -dp_directional<dpsite>(dir))), @@ -755,27 +759,27 @@ void init() { // FIXME: border::adjust(input, win.delta()); - extension::fill(input, mln_max(mln_value(I))); + extension::fill(input, accu); initialize(output, input); } void next() { for_all(q_l) - min.untake(q_l.val()); + accu.untake(q_l.val()); for_all(q_r) - min.take(q_r.val()); - output(p) = min; + accu.take(q_r.val()); + output(p) = accu; } void init_run() { - min.init(); + accu.init(); p[dir]--; mln_qixter(const I, W) q(input, win, p); for_all(q) - min.take(q.val()); + accu.take(q.val()); p[dir]++; } @@ -793,7 +797,11 @@ { trace::entering("morpho::impl:erosion_directional_nd_fastest"); - typedef erosion_directional_nd_fastest_functor<I, W> F; + typedef mlc_is(mln_trait_image_kind(I), + trait::image::kind::binary) is_binary; + typedef mlc_if(is_binary, accu::land, accu::min_h<mln_value(I)>) A; + + typedef erosion_directional_nd_fastest_functor<I, W, A> F; F f(exact(input), exact(win), dir); canvas::browsing::directional(f); @@ -1400,13 +1408,10 @@ return erosion_dispatch_for_generic(input, win); else { - typedef mlc_is_not(mln_trait_image_kind(I), - mln::trait::image::kind::logic) test_not_logic; typedef mlc_is_a(mln_pset(I), Box) test_box; typedef mlc_equal(mln_trait_image_quant(I), mln::trait::image::quant::low) test_lowq; - typedef mlc_and(test_not_logic, test_box) temp; - typedef mlc_and(temp, test_lowq) tests; + typedef mlc_and(test_box, test_lowq) tests; return erosion_dispatch_wrt_win(typename tests::eval (), input, win); } @@ -1445,13 +1450,10 @@ return erosion_dispatch_for_generic(input, win); else { - typedef mlc_is_not(mln_trait_image_kind(I), - mln::trait::image::kind::logic) test_not_logic; typedef mlc_is_a(mln_pset(I), Box) test_box; typedef mlc_equal(mln_trait_image_quant(I), mln::trait::image::quant::low) test_lowq; - typedef mlc_and(test_not_logic, test_box) temp; - typedef mlc_and(temp, test_lowq) tests; + typedef mlc_and(test_box, test_lowq) tests; return erosion_dispatch_wrt_win(typename tests::eval (), input, win); } Index: mln/morpho/includes.hh --- mln/morpho/includes.hh (revision 2667) +++ mln/morpho/includes.hh (working copy) @@ -44,6 +44,8 @@ # include <mln/value/ops.hh> +# include <mln/accu/land.hh> +// # include <mln/accu/lor.hh> # include <mln/accu/min.hh> # include <mln/accu/max.hh> # include <mln/accu/min_h.hh>

1 0

2667: Fix an endless loop.
by Guillaume Lazzara 23 Oct '08

23 Oct '08

* scribo/demat.hh: Fix a endless loop. --- milena/sandbox/ChangeLog | 6 +++ milena/sandbox/scribo/demat.hh | 95 +++++++++++++++++++++++++++++++++++----- 2 files changed, 90 insertions(+), 11 deletions(-) diff --git a/milena/sandbox/ChangeLog b/milena/sandbox/ChangeLog index 512eac2..9faa0a9 100644 --- a/milena/sandbox/ChangeLog +++ b/milena/sandbox/ChangeLog @@ -1,3 +1,9 @@ +2008-10-23 Guillaume Lazzara <z(a)lrde.epita.fr> + + Fix an endless loop. + + * scribo/demat.hh: Fix a endless loop. + 2008-10-22 Maxime van Noppen <yabo(a)lrde.epita.fr> Work on ICCVG04 for INIM. diff --git a/milena/sandbox/scribo/demat.hh b/milena/sandbox/scribo/demat.hh index a6985d5..5f01993 100644 --- a/milena/sandbox/scribo/demat.hh +++ b/milena/sandbox/scribo/demat.hh @@ -100,6 +100,13 @@ namespace scribo using value::rgb8; + void draw_component_boxes(image2d<rgb8>& output, const util::array<box2d>& boxes) + { + for (unsigned i = 1; i < boxes.nelements(); ++i) + if (boxes[i].is_valid()) + draw::box(output, boxes[i], literal::red); + } + template <typename V> void save_lbl_image(const image2d<V>& lbl, unsigned nlabels, const char *filename) @@ -126,6 +133,14 @@ namespace scribo return level::transform(input, fconvert); } + void negate(image2d<bool>& in) + { + fun::i2v::array<unsigned> negate(2); + negate(0) = 1; + negate(1) = 0; + level::apply(in, negate); + } + /// Functions related to the table removal /// \{ @@ -207,8 +222,8 @@ namespace scribo #ifndef NOUT image2d<rgb8> tmp = clone(output); - erase_table_boxes(tmp, vboxes, n); - erase_table_boxes(tmp, hboxes, n); + draw_component_boxes(tmp, vboxes); + draw_component_boxes(tmp, hboxes); io::ppm::save(tmp, "./table-filtered.ppm"); #endif @@ -222,6 +237,7 @@ namespace scribo /// Function related to text extraction /// \{ + inline int_u16 most_left(const fun::i2v::array<int_u16>& left_link, unsigned i) { @@ -230,6 +246,15 @@ namespace scribo return i; } + inline + int_u16 + uncurri_left_link(const fun::i2v::array<int_u16>& left_link, unsigned i) + { + if (left_link(i) != i) + left_link(i) = most_left(left_link, left_link(i)); + return left_link(i); + } + template <typename V> void remove_small_comps_i2v(image2d<V>& lbl, @@ -276,12 +301,14 @@ namespace scribo tboxes.resize(ncomp + 1); for (unsigned i = 1; i <= ncomp; ++i) { - if (left_link(i) != i) - left_link(i) = most_left(left_link, left_link(i)); - tboxes[left_link(i)].take(cboxes[i]); + //if (left_link(i) != i) + // left_link(i) = most_left(left_link, left_link(i)); + tboxes[most_left(left_link, i)].take(cboxes[i]); } //Update labels + for (unsigned i = 1; i <= ncomp; ++i) + uncurri_left_link(left_link, i); level::apply(lbl, left_link); #ifndef NOUT @@ -290,7 +317,8 @@ namespace scribo util::array<box2d> result; for (unsigned i = 1; i <= ncomp; ++i) - result.append(tboxes[i].to_result()); + if (tboxes[i].is_valid()) + result.append(tboxes[i].to_result()); return result; } @@ -305,7 +333,7 @@ namespace scribo if (lbl.domain().has(p) && lbl(p) != 0 && lbl(p) != i && (p.col() - c.col()) < dmax) { - if (left_link(lbl(p)) == lbl(p)) + if (left_link(lbl(p)) == lbl(p) && most_left(left_link, i) != lbl(p)) left_link(lbl(p)) = i; // else // left_link(lbl(p)) = 0;//FIXME: should be uncommented? @@ -350,7 +378,39 @@ namespace scribo return left_link; } +/* + void merge_bboxes(util::array<box2d>& cboxes, + image2d<int_u16>& lbl, unsigned &nlabels) + { + fun::i2v::array<int_u16> merge; + unsigned current = 1; + for (unsigned i = 1; i <= nlabels;) + { + unsigned midcol = (cboxes[i].pmax().col() - cboxes[i].pmin().col()) / 2; + /// FIXME: Box center => Routine? + point2d c (cboxes[i].pmin().row() + + ((cboxes[i].pmax().row() - cboxes[i].pmin().row()) / 2), + cboxes[i].pmin().col() + midcol); + /// First site on the right of the center site + point2d p(c.row(), c.col() + 1); + + // FIXME: Lemmings with a condition on the distance => write a special version? + while (lbl.domain().has(p) && (lbl(p) == 0 || lbl(p) == i) + && (p.col() - c.col()) <= midcol) + ++p.col(); + + if (lbl.domain().has(p) && lbl(p) != 0 && lbl(p) != i + && (p.col() - c.col()) <= midcol) + { + + } + } + + level::apply(lbl, merge); + cboxes.resize(nlabels); + } +*/ util::array<box2d> extract_text(image2d<bool>& in, @@ -373,6 +433,14 @@ namespace scribo boxes_t cboxes = labeling::compute(accu::meta::bbox(), lbl, nlabels); +#ifndef NOUT + image2d<rgb8> tmp = clone(output); + draw_component_boxes(tmp, cboxes); + io::ppm::save(tmp, "character-bboxes.ppm"); +#endif + + //merge_bboxes(cboxes, lbl, nlabels); + //Link character bboxes to their left neighboor if possible. fun::i2v::array<int_u16> left = link_character_bboxes(lbl, cboxes, nlabels, bbox_distance); @@ -407,6 +475,12 @@ namespace scribo //Load image image2d<bool> in; io::pbm::load(in, argv[1]); + internal::negate(in); + +#ifndef NOUT + io::pbm::save(in, "in-neg.pbm"); +#endif + image2d<rgb8> output = internal::to_color(in); internal::extract_tables(in, output, l, bbox_larger); @@ -414,13 +488,12 @@ namespace scribo util::array<box2d> tboxes = internal::extract_text(in, output, bbox_distance, min_comp_size); - for (unsigned i = 1; i < tboxes.nelements(); ++i) - if (tboxes[i].is_valid()) - draw::box(output, tboxes[i], literal::red); + internal::draw_component_boxes(output, tboxes); io::ppm::save(output, argv[2]); /// Use txt bboxes here with Tesseract - /// => in | tboxes[i] + /// for (i = 1; i < tboxes.nelements(); ++i) + /// tesseract(in | tboxes[i]) } } // end of namespace scribo -- 1.5.6.5

1 0

2666: Improve mln::io::off:load.
by Roland Levillain 23 Oct '08

23 Oct '08

* mln/io/off/load.hh (mln::io::off:load): Ensure the end of the file is reached. --- milena/ChangeLog | 7 +++++++ milena/mln/io/off/load.hh | 11 +++++++++++ 2 files changed, 18 insertions(+), 0 deletions(-) diff --git a/milena/ChangeLog b/milena/ChangeLog index 4d2c9e9..2c8f00d 100644 --- a/milena/ChangeLog +++ b/milena/ChangeLog @@ -1,5 +1,12 @@ 2008-10-23 Roland Levillain <roland(a)lrde.epita.fr> + Improve mln::io::off:load. + + * mln/io/off/load.hh (mln::io::off:load): Ensure the end of the + file is reached. + +2008-10-23 Roland Levillain <roland(a)lrde.epita.fr> + Adjust I/O routines to algebraic faces. * mln/io/off/load.hh (mln::io::off::load): Adjust. diff --git a/milena/mln/io/off/load.hh b/milena/mln/io/off/load.hh index 38e6b6e..60c534b 100644 --- a/milena/mln/io/off/load.hh +++ b/milena/mln/io/off/load.hh @@ -264,6 +264,17 @@ namespace mln // Image. ima.init_(pc, values); + /*--------------. + | End of file. | + `--------------*/ + + istr >> internal::eat_comment; + if (!istr.eof()) + { + std::cerr << me << ": `" << filename + << "': end of file not reached" << std::endl; + std::exit(1); + } istr.close(); } -- 1.5.6.5

1 0

2665: Adjust I/O routines to algebraic faces.
by Roland Levillain 23 Oct '08

23 Oct '08

* mln/io/off/load.hh (mln::io::off::load): Adjust. --- milena/ChangeLog | 6 ++++++ milena/mln/io/off/load.hh | 6 ++++-- 2 files changed, 10 insertions(+), 2 deletions(-) diff --git a/milena/ChangeLog b/milena/ChangeLog index e32143d..4d2c9e9 100644 --- a/milena/ChangeLog +++ b/milena/ChangeLog @@ -1,5 +1,11 @@ 2008-10-23 Roland Levillain <roland(a)lrde.epita.fr> + Adjust I/O routines to algebraic faces. + + * mln/io/off/load.hh (mln::io::off::load): Adjust. + +2008-10-23 Roland Levillain <roland(a)lrde.epita.fr> + Convert iterators on complexes to algebraic faces. * mln/topo/internal/complex_relative_iterator_base.hh diff --git a/milena/mln/io/off/load.hh b/milena/mln/io/off/load.hh index 6cdf3d5..38e6b6e 100644 --- a/milena/mln/io/off/load.hh +++ b/milena/mln/io/off/load.hh @@ -220,14 +220,16 @@ namespace mln topo::n_face<0, D> vertex(c, vertex_id); topo::n_face<0, D> next_vertex(c, next_vertex_id); // The current edge. - topo::n_face<1, D> edge; + topo::algebraic_n_face<1, D> edge; // If the edge has been constructed yet, create it; // otherwise, retrieve its id from the complex. if (!complex_edges[vertex_id][next_vertex_id]) { complex_edges[vertex_id][next_vertex_id] = true; complex_edges[next_vertex_id][vertex_id] = true; - edge = c.add_face(vertex + next_vertex); + edge = + make_algebraic_n_face(c.add_face(vertex - next_vertex), + true); } else { -- 1.5.6.5

1 0