October 2012 - Olena-patches

[Olena] #244: Make use of \snippet command with doxygen
by Olena Trac 29 Nov '12

29 Nov '12

#244: Make use of \snippet command with doxygen -------------------------+-------------------------------------------------- Reporter: lazzara | Owner: Olena Team Type: enhancement | Status: new Priority: major | Milestone: Olena 1.2 Component: Milena | Version: 1.0 Keywords: | -------------------------+-------------------------------------------------- Sample codes inserted in the Doxygen documentation are currently extracted from full source codes using a shell script. It enables the possibility to check code compilation and give full working code to the users. Since version 1.7.5, Doxygen added a new command, \snippet, which has this functionality of splitting full code examples easily. We should use it and remove all the current scripts. -- Ticket URL: <https://trac.lrde.org/olena/ticket/244> Olena <http://olena.lrde.epita.fr> Olena, a generic and efficient C++ image processing library.

1 3

[Olena] #218: Check for unnecessary includes.
by Olena Trac 29 Nov '12

29 Nov '12

#218: Check for unnecessary includes. ---------------------+------------------------------------------------------ Reporter: lazzara | Owner: Olena Team Type: defect | Status: new Priority: major | Milestone: Olena 1.1 Component: Milena | Version: 1.0 Keywords: | ---------------------+------------------------------------------------------ Few files may includes unnecessary files. Write a program and include these files: #include <mln/core/var.hh> #include <mln/core/image/image2d.hh> #include <mln/debug/iota.hh> #include <mln/debug/println.hh> #include <mln/pw/image.hh> You will be able to use p_array and sub_image without including their related files. Unnecessary includes slow down compilation and is damaging. -- Ticket URL: <https://trac.lrde.org/olena/ticket/218> Olena <http://olena.lrde.epita.fr> Olena, a generic and efficient C++ image processing library.

1 1

[Olena] #195: Organize accumulators in subdirectories
by Olena Trac 29 Nov '12

29 Nov '12

#195: Organize accumulators in subdirectories -------------------------+-------------------------------------------------- Reporter: theo | Owner: Olena Team Type: enhancement | Status: new Priority: critical | Milestone: Olena 1.0 Component: Milena | Version: 1.0 Keywords: | -------------------------+-------------------------------------------------- For instance: * `accu/stats/` * ... * What else? -- Ticket URL: <https://trac.lrde.org/olena/ticket/195> Olena <http://olena.lrde.epita.fr> Olena, a generic and efficient C++ image processing library.

1 3

[Olena] #251: Document how to compile with multiple files
by Olena Trac 29 Nov '12

29 Nov '12

#251: Document how to compile with multiple files -------------------------+-------------------------------------------------- Reporter: lazzara | Owner: lazzara Type: enhancement | Status: new Priority: minor | Milestone: Olena 2.1 Component: Milena | Version: 2.0 Keywords: | -------------------------+-------------------------------------------------- Add a chapter in the user reference guide. - MLN_INCLUDE_ONLY - MLN_WO_GLOBAL_VARS - Compilation with automake - Compilation with QMake - Common error messages (duplicated symbols, undefined symbols) -- Ticket URL: <https://trac.lrde.org/olena/ticket/251> Olena <http://olena.lrde.epita.fr> Olena, a software platform dedicated to image processing.

1 2

[Olena] #208: Fix dependencies of the HTML documentation generation?
by Olena Trac 29 Nov '12

29 Nov '12

#208: Fix dependencies of the HTML documentation generation? ----------------------+----------------------------------------------------- Reporter: levill_r | Owner: Olena Team Type: defect | Status: new Priority: minor | Milestone: Olena 1.1 Component: Milena | Version: 1.0 Keywords: doc | ----------------------+----------------------------------------------------- The generation of the HTML documentation requires a file named `html.sty`, which is not provided by HeVeA, but by latex2html. Is that right? I thought we only had a dependency on the former, not the latter. -- Ticket URL: <https://trac.lrde.org/olena/ticket/208> Olena <http://olena.lrde.epita.fr> Olena, a generic and efficient C++ image processing library.

1 1

[Olena] #219: Add \internal in the documentation of internal code
by Olena Trac 29 Nov '12

29 Nov '12

#219: Add \internal in the documentation of internal code ---------------------+------------------------------------------------------ Reporter: lazzara | Owner: Olena Team Type: defect | Status: new Priority: major | Milestone: Olena 1.1 Component: Milena | Version: 1.0 Keywords: | ---------------------+------------------------------------------------------ Currently, in order to not expose the internal code in the user documentation, everything in the internal and trait namespaces is ignored during the doc generation. The main issue is that inherited members are not shown in the documentation of the classes exposed to the user. The only way to ask Doxygen to show inherited members and not to show internal documentation is to add \internal in the documentation of internal code. -- Ticket URL: <https://trac.lrde.org/olena/ticket/219> Olena <http://olena.lrde.epita.fr> Olena, a generic and efficient C++ image processing library.

1 2

[Olena] #235: Support GraphicsMagick++ and drop support for ImageMagick++
by Olena Trac 29 Nov '12

29 Nov '12

#235: Support GraphicsMagick++ and drop support for ImageMagick++ -------------------------+-------------------------------------------------- Reporter: levill_r | Owner: Olena Team Type: proposal | Status: new Priority: major | Milestone: Olena 1.1 Component: Milena | Version: 1.0 Keywords: io external | -------------------------+-------------------------------------------------- !GraphicsMagick (http://www.graphicsmagick.org/) seems to provide a framework more modern than !ImageMagick's (http://www.imagemagick.org) in particular regarding multi-core support. Moreover Fabien has discovered weird behaviors (#210) and Buildbot has also recently exhibited run time errors that did not surface before (https://buildfarm.lrde.org/buildfarm/oln/builders/i486-linux-gnu- gcc-4.1/builds/539/steps/check/logs/stdio) -- though I don't know who's the culprit here. -- Ticket URL: <https://trac.lrde.org/olena/ticket/235> Olena <http://olena.lrde.epita.fr> Olena, a generic and efficient C++ image processing library.

1 2

branch exp/khalimsky updated: olena-2.0-166-gf550cfa
by git＠lrde.epita.fr 31 Oct '12

31 Oct '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "Olena, a generic and efficient image processing platform". The branch exp/khalimsky has been updated via f550cfab304c3851c40b188029790114f74c17de (commit) from 4821f6f79b7795c3e30cf1bf20669d6160bec480 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- f550cfa mln/world/kn/display_enlarged.hh: Fix. ----------------------------------------------------------------------- Summary of changes: milena/ChangeLog | 4 ++++ milena/mln/world/kn/display_enlarged.hh | 30 ++++++++++++++++-------------- 2 files changed, 20 insertions(+), 14 deletions(-) hooks/post-receive -- Olena, a generic and efficient image processing platform

1 0

olena-2.0-166-gf550cfa mln/world/kn/display_enlarged.hh: Fix.
by Guillaume Lazzara 31 Oct '12

31 Oct '12

--- milena/ChangeLog | 4 ++++ milena/mln/world/kn/display_enlarged.hh | 30 ++++++++++++++++-------------- 2 files changed, 20 insertions(+), 14 deletions(-) diff --git a/milena/ChangeLog b/milena/ChangeLog index 475fb8e..6dedd0d 100644 --- a/milena/ChangeLog +++ b/milena/ChangeLog @@ -1,3 +1,7 @@ +2012-10-31 Guillaume Lazzara <z(a)lrde.epita.fr> + + * mln/world/kn/display_enlarged.hh: Fix. + 2012-10-30 Guillaume Lazzara <z(a)lrde.epita.fr> * mln/world/k1/immerse.hh: Fix call to an overload in kn diff --git a/milena/mln/world/kn/display_enlarged.hh b/milena/mln/world/kn/display_enlarged.hh index 5d1ef50..e1b122a 100644 --- a/milena/mln/world/kn/display_enlarged.hh +++ b/milena/mln/world/kn/display_enlarged.hh @@ -93,16 +93,18 @@ namespace mln mln_assertion(zoom_ % 2); const I& ima_kn = exact(ima_kn_); - int shift = zoom_ - 1; + int + shift = zoom_ + 1, + mult = shift / 2; // No zoom: return original image. if (zoom_ == 1) - return ima_kn; + return duplicate(ima_kn); - mln_concrete(I) output(make::box2d(ima_kn.domain().pmin().row() * shift, - ima_kn.domain().pmin().col() * shift, - ima_kn.domain().pmax().row() * shift, - ima_kn.domain().pmax().col() * shift)); + mln_concrete(I) output(make::box2d(ima_kn.domain().pmin().row() * mult, + ima_kn.domain().pmin().col() * mult, + ima_kn.domain().pmax().row() * mult, + ima_kn.domain().pmax().col() * mult)); const mln::def::coord min_row = geom::min_row(ima_kn), @@ -116,23 +118,23 @@ namespace mln { if (kn::is_0_face(row, col)) { - output.at_(row * shift, col * shift) = ima_kn.at_(row, col); + output.at_(row * mult, col * mult) = ima_kn.at_(row, col); } else if (kn::is_1_face_vertical(row, col)) { - for (mln::def::coord j = -(shift - 1); j < (shift + 3); ++j) - output.at_(row * shift + j, col * shift) = ima_kn.at_(row, col); + for (mln::def::coord j = -(mult - 1); j <= (mult - 1); ++j) + output.at_(row * mult + j, col * mult) = ima_kn.at_(row, col); } else if (kn::is_1_face_horizontal(row, col)) { - for (mln::def::coord j = -(shift - 1); j < (shift + 3); ++j) - output.at_(row * shift, col * shift + j) = ima_kn.at_(row, col); + for (mln::def::coord j = -(mult - 1); j <= (mult - 1); ++j) + output.at_(row * mult, col * mult + j) = ima_kn.at_(row, col); } else if (kn::is_2_face(row, col)) { - for (mln::def::coord i = -(shift - 1); i < (shift + 3); ++i) - for (mln::def::coord j = -(shift - 1); j < (shift + 3); ++j) - output.at_(row * shift + i, col * shift + j) = ima_kn.at_(row, col); + for (mln::def::coord i = -(mult - 1); i <= (mult - 1); ++i) + for (mln::def::coord j = -(mult - 1); j <= (mult - 1); ++j) + output.at_(row * mult + i, col * mult + j) = ima_kn.at_(row, col); } } } -- 1.7.2.5

1 0

olena-2.0-28-geeecbfb fix reading and writing the same image using neighbours
by Coddy Levi 30 Oct '12

30 Oct '12

--- milena/mln/inpainting/lpde.hh | 67 +++++++++++++++++++++++++---------------- 1 files changed, 41 insertions(+), 26 deletions(-) diff --git a/milena/mln/inpainting/lpde.hh b/milena/mln/inpainting/lpde.hh index 04a903a..44d036e 100644 --- a/milena/mln/inpainting/lpde.hh +++ b/milena/mln/inpainting/lpde.hh @@ -94,8 +94,9 @@ namespace mln } template <typename I, typename M> - float diffuse(I& u, - const I& f, + float diffuse(const I& u, + I& u_next, + //const I& f, const std::vector<float>& w, const M& mask) { @@ -105,7 +106,7 @@ namespace mln const unsigned derivative_kernels_size = 5; const float* derivative_kernel[derivative_kernels_size]; - + derivative_kernel[0] = make_dx_kernel(); derivative_kernel[1] = make_dy_kernel(); derivative_kernel[2] = make_dxx_kernel(); @@ -119,25 +120,26 @@ namespace mln convert::from_to(vals, full3x3); - mln_pixter(I) p(u); + mln_pixter(const I) p(u); + mln_pixter(I) p_next(u_next); //mln_pixter(const I) p_f(f); mln_pixter(const M) p_m(mask); - mln_qixter(const I, window2d) q(p, full3x3); + mln_qixter(const I, const window2d) q(p, full3x3); float sse = 0; - + unsigned car = 0; - for_all_2 (p, p_m) - { - //if (p_m.val()) + for_all_3 (p, p_next, p_m) + { + //if (p_m.val()) { - T old = p.val(); + const T old = p.val(); if (fabs(w[1]) > activation_threshold) - p.val() *= w[1]; + p_next.val() = p.val() * w[1]; // if (fabs(w[0]) > activation_threshold) - // p.val() += w[0] * p_f.val(); + // p_next.val() += w[0] * p_f.val(); T p_derivative[derivative_kernels_size]; @@ -168,25 +170,25 @@ namespace mln //||∇u||^2 if (fabs(w[2]) > activation_threshold) - p.val() += w[2] * (p_x2 + p_y2); + p_next.val() += w[2] * (p_x2 + p_y2); // laplacian if (fabs(w[3]) > activation_threshold) - p.val() += w[3] * (p_xx + p_yy); + p_next.val() += w[3] * (p_xx + p_yy); // tr(H^2(u)) if (fabs(w[4]) > activation_threshold) - p.val() += w[4] * (p_xx2 + 2 * p_xy2 + p_yy2); + p_next.val() += w[4] * (p_xx2 + 2 * p_xy2 + p_yy2); // (∇u)^T H_u (∇u) if (fabs(w[5]) > activation_threshold) - p.val() += w[5] * (components_product(p_x2, p_xx) - + 2 * components_product(components_product(p_x, p_y), p_xy) - + components_product(p_y2, p_yy)); + p_next.val() += w[5] * (components_product(p_x2, p_xx) + + 2 * components_product(components_product(p_x, p_y), p_xy) + + components_product(p_y2, p_yy)); - sse += norm::l2(old - p.val()); + sse += norm::l2(old - p_next.val()); } - } + } return sse; } @@ -195,18 +197,31 @@ namespace mln void lpde::operator()(I<T>& src, I<bool>& mask) { - extension::adjust_duplicate(src, 1); + //const I<T> f(src); + + I<T>& u_ = src; + I<T> u_next_(src); + + extension::adjust_duplicate(u_, 1); + extension::adjust_duplicate(u_next_, 1); - const I<T> f(src); + I<T>* u = &u_; + I<T>* u_next = &u_next_; - I<T>& u = src; + float sse = std::numeric_limits<float>::max() / 2.0; - float sse = std::numeric_limits<float>::max(); + float sse_old = std::numeric_limits<float>::max(); float t = 0; - while (t < this->t_f && sse > 0.01 && !std::isnan(sse)) + while (t < this->t_f && sse < sse_old && !std::isnan(sse)) { - sse = diffuse(u, f, w, mask); + sse_old = sse; + + sse = diffuse(*u, *u_next, /* f, */ w, mask); + + I<T>* swap = u; + u = u_next; + u_next = swap; std::cout << "sse = " << sse << std::endl; -- 1.7.2.5

1 0