- Olena-patches

olena-2.0-385-g62dbacd Remove references to external directory.
by Guillaume Lazzara 13 Mar '13

13 Mar '13

* configure.ac, * Makefile.am: Here. --- ChangeLog | 7 +++++++ Makefile.am | 2 +- configure.ac | 1 - 3 files changed, 8 insertions(+), 2 deletions(-) diff --git a/ChangeLog b/ChangeLog index 5cc8a6c..0754452 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,10 @@ +2013-03-13 Guillaume Lazzara <z(a)lrde.epita.fr> + + Remove references to external directory. + + * configure.ac, + * Makefile.am: Here. + 2013-03-08 Guillaume Lazzara <z(a)lrde.epita.fr> Improve look'n feel of general documentation. diff --git a/Makefile.am b/Makefile.am index 2f3ba94..b14438e 100644 --- a/Makefile.am +++ b/Makefile.am @@ -17,7 +17,7 @@ ACLOCAL_AMFLAGS = -I m4 -SUBDIRS = build-aux external milena +SUBDIRS = build-aux milena if ENABLE_SWILENA SUBDIRS += swilena diff --git a/configure.ac b/configure.ac index 2f96ca2..3b5a3aa 100644 --- a/configure.ac +++ b/configure.ac @@ -414,7 +414,6 @@ AC_CONFIG_FILES([milena/doc/tools/data.hh]) AC_CONFIG_FILES([ Makefile doc/Makefile - external/Makefile milena/Makefile milena/doc/Makefile milena/mesh/Makefile -- 1.7.2.5

1 0

olena-2.0-401-g1821109 Remove references to external directory.
by Guillaume Lazzara 13 Mar '13

13 Mar '13

* configure.ac, * Makefile.am: Here. Signed-off-by: Roland Levillain <roland(a)lrde.epita.fr> --- ChangeLog | 7 +++++++ Makefile.am | 2 +- configure.ac | 1 - 3 files changed, 8 insertions(+), 2 deletions(-) diff --git a/ChangeLog b/ChangeLog index 5cc8a6c..0754452 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,10 @@ +2013-03-13 Guillaume Lazzara <z(a)lrde.epita.fr> + + Remove references to external directory. + + * configure.ac, + * Makefile.am: Here. + 2013-03-08 Guillaume Lazzara <z(a)lrde.epita.fr> Improve look'n feel of general documentation. diff --git a/Makefile.am b/Makefile.am index 2f3ba94..b14438e 100644 --- a/Makefile.am +++ b/Makefile.am @@ -17,7 +17,7 @@ ACLOCAL_AMFLAGS = -I m4 -SUBDIRS = build-aux external milena +SUBDIRS = build-aux milena if ENABLE_SWILENA SUBDIRS += swilena diff --git a/configure.ac b/configure.ac index 2f96ca2..3b5a3aa 100644 --- a/configure.ac +++ b/configure.ac @@ -414,7 +414,6 @@ AC_CONFIG_FILES([milena/doc/tools/data.hh]) AC_CONFIG_FILES([ Makefile doc/Makefile - external/Makefile milena/Makefile milena/doc/Makefile milena/mesh/Makefile -- 1.7.2.5

1 0

olena-2.0-389-g277ebe7c Start new toolchain (scribo toolchain inspired)
by Anthony Seure 12 Mar '13

12 Mar '13

--- scribo/sandbox/icdar_13_table/Makefile | 2 +- scribo/sandbox/icdar_13_table/src/new.cc | 146 ++++++++++++++++++++++++++++++ 2 files changed, 147 insertions(+), 1 deletions(-) create mode 100644 scribo/sandbox/icdar_13_table/src/new.cc diff --git a/scribo/sandbox/icdar_13_table/Makefile b/scribo/sandbox/icdar_13_table/Makefile index 4e552dd..b9600a4 100644 --- a/scribo/sandbox/icdar_13_table/Makefile +++ b/scribo/sandbox/icdar_13_table/Makefile @@ -4,7 +4,7 @@ CFLAGS=-Wall -Werror -O3 -DHAVE_TESSERACT_3 -DNDEBUG CLIBS=-I../../../milena/ -I../../ -I/usr/include/poppler CLEAN=*.o output/* log final.xml -SRC=src/main.cc +SRC=src/new.cc OUTPUT=table all: table diff --git a/scribo/sandbox/icdar_13_table/src/new.cc b/scribo/sandbox/icdar_13_table/src/new.cc new file mode 100644 index 0000000..963aa7d --- /dev/null +++ b/scribo/sandbox/icdar_13_table/src/new.cc @@ -0,0 +1,146 @@ +#include <mln/binarization/all.hh> + +#include <mln/core/image/image2d.hh> + +#include <mln/data/all.hh> +#include <mln/draw/line.hh> + +#include <mln/fun/v2v/rgb_to_luma.hh> + +#include <mln/io/pbm/all.hh> +#include <mln/io/pdf/load.hh> +#include <mln/io/ppm/all.hh> + +#include <mln/labeling/all.hh> +#include <mln/literal/all.hh> +#include <mln/logical/and.hh> + +#include <mln/value/all.hh> + +#include <tesseract/baseapi.h> + +#include <scribo/binarization/sauvola.hh> +#include <scribo/core/component_set.hh> +#include <scribo/preprocessing/denoise_fg.hh> +#include <scribo/primitive/extract/vertical_separators.hh> +#include <scribo/primitive/remove/separators.hh> +#include <scribo/primitive/extract/separators_nonvisible.hh> + +#include <scribo/primitive/link/internal/dmax_width_and_height.hh> +#include <scribo/primitive/link/with_single_right_link_dmax_ratio.hh> +#include <scribo/primitive/link/with_single_left_link_dmax_ratio.hh> +#include <scribo/primitive/link/merge_double_link.hh> + +using namespace mln; + +void start_xml(std::ofstream& xml, const char* name, const char* pdf) +{ + xml.open(name); + xml << "<?xml version\"1.0\" encoding=\"UTF-8\"?>" << std::endl + << "<document filename='" << pdf << "'>" << std::endl; +} + +void end_xml(std::ofstream& xml) +{ + xml << "</document>" << std::endl; + xml.close(); +} + +void write_table(std::ofstream& xml, const point2d& start, const point2d& end) +{ + static unsigned table = 0; + static unsigned region = 0; + static unsigned page = 1; + + xml << "\t<table id='" << table << "'>" << std::endl + << "\t\t<region id='" << region << "' page='" << page << "'>" << std::endl + << "\t\t<bounding-box x1='" << start[1] << "' y1='" << start[0] << "' " + << "x2='" << end[1] << "' y2='" << end[0] << "'/>" << std::endl + << "\t\t</region>" << std::endl + << "\t</table>" << std::endl; + + ++table; +} + +int main(int argc, char** argv) +{ + typedef value::label_16 V; + typedef image2d<V> L; + + std::ofstream xml; + std::ostringstream path; + image2d<value::rgb8> original; + image2d<value::int_u8> filtered; + image2d<bool> bin, separators, bin_without_separators, whitespaces, + denoised, comp, links; + scribo::component_set< image2d<unsigned> > components; + + unsigned dpi = 72; + + // Loading and binarization + start_xml(xml, "final.xml", argv[1]); + + util::array< image2d<value::rgb8> > pdf; + io::pdf::load(pdf, argv[1], dpi); + for (unsigned page = 0; page < pdf.nelements(); ++page) + { + original = pdf[page]; + filtered = data::transform(original, fun::v2v::rgb_to_luma<value::int_u8>()); + bin = scribo::binarization::sauvola(filtered, 81, 0.44); + + // Find separators + separators = scribo::primitive::extract::vertical_separators(bin, 81); + bin_without_separators = scribo::primitive::remove::separators(bin, separators); + whitespaces = scribo::primitive::extract::separators_nonvisible(bin); + + // Denoise + denoised = scribo::preprocessing::denoise_fg(bin_without_separators, c8(), 4); + + // Extract components + unsigned ncomponents; + components = scribo::primitive::extract::components(denoised, c8(), ncomponents); + + initialize(comp, denoised); + data::fill(comp, false); + for (unsigned i = 1; i <= components.nelements(); ++i) + data::fill((comp | components(i).bbox()).rw(), true); + + scribo::object_links< image2d<unsigned> > right_link = scribo::primitive::link::with_single_right_link_dmax_ratio(components, + scribo::primitive::link::internal::dmax_width_and_height(1), + scribo::anchor::MassCenter); + + scribo::object_links< image2d<unsigned> > left_link = scribo::primitive::link::with_single_left_link_dmax_ratio(components, + scribo::primitive::link::internal::dmax_width_and_height(1), + scribo::anchor::MassCenter); + + scribo::object_links< image2d<unsigned> > merged_links = scribo::primitive::link::merge_double_link(left_link, right_link); + + initialize(links, denoised); + data::fill(links, false); + for (unsigned i = 1; i <= merged_links.components().nelements(); ++i) + data::fill((links | merged_links.components()(i).bbox()).rw(), true); + + // Write images and close XML + path.str(""); path << "output/p" << page << "_0_bin.pbm"; + io::pbm::save(bin, path.str()); + + path.str(""); path << "output/p" << page << "_1_bin_without_separators.pbm"; + io::pbm::save(bin_without_separators, path.str()); + + path.str(""); path << "output/p" << page << "_2_whitespaces.pbm"; + io::pbm::save(whitespaces, path.str()); + + path.str(""); path << "output/p" << page << "_3_denoised.pbm"; + io::pbm::save(denoised, path.str()); + + path.str(""); path << "output/p" << page << "_4_components.pbm"; + io::pbm::save(comp, path.str()); + + path.str(""); path << "output/p" << page << "_5_links.pbm"; + io::pbm::save(links, path.str()); + } + + end_xml(xml); + + return 0; +} -- 1.7.2.5

1 0

branch sandbox/anthony updated: olena-2.0-388-gd4b28f4
by git＠lrde.epita.fr 11 Mar '13

11 Mar '13

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 sandbox/anthony has been updated via d4b28f4648097e6b349300f4fb4d89449629f490 (commit) from 99b36312a31ce549eb18103576794165e61281f3 (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 ----------------------------------------------------------------- d4b28f4 Work with mutlipages PDF ----------------------------------------------------------------------- Summary of changes: scribo/sandbox/icdar_13_table/Makefile | 13 ++- scribo/sandbox/icdar_13_table/src/main.cc | 184 +++++++++++++++++------------ 2 files changed, 114 insertions(+), 83 deletions(-) hooks/post-receive -- Olena, a generic and efficient image processing platform

1 0

olena-2.0-388-gd4b28f4 Work with mutlipages PDF
by Anthony Seure 11 Mar '13

11 Mar '13

--- scribo/sandbox/icdar_13_table/Makefile | 13 ++- scribo/sandbox/icdar_13_table/src/main.cc | 184 +++++++++++++++++------------ 2 files changed, 114 insertions(+), 83 deletions(-) diff --git a/scribo/sandbox/icdar_13_table/Makefile b/scribo/sandbox/icdar_13_table/Makefile index 2420eec..4e552dd 100644 --- a/scribo/sandbox/icdar_13_table/Makefile +++ b/scribo/sandbox/icdar_13_table/Makefile @@ -1,8 +1,8 @@ CCACHE= CC=g++ -CFLAGS=-Wall -Werror -O3 -CLIBS=-I../../../milena/ -I../../ -CLEAN=*.o $(OUTPUT) output/* log final.xml +CFLAGS=-Wall -Werror -O3 -DHAVE_TESSERACT_3 -DNDEBUG +CLIBS=-I../../../milena/ -I../../ -I/usr/include/poppler +CLEAN=*.o output/* log final.xml SRC=src/main.cc OUTPUT=table @@ -10,9 +10,12 @@ OUTPUT=table all: table table: - $(CCACHE) $(CC) $(CFLAGS) -DHAVE_TESSERACT_3 $(CLIBS) $(SRC) -ltesseract -o $(OUTPUT) + $(CCACHE) $(CC) $(CFLAGS) $(CLIBS) $(SRC) -ltesseract -lpoppler-cpp -o $(OUTPUT) clean: rm -rf $(CLEAN) -.PHONY: table +mrproper: clean + rm -f $(OUTPUT) + +.PHONY: table clean mrproper diff --git a/scribo/sandbox/icdar_13_table/src/main.cc b/scribo/sandbox/icdar_13_table/src/main.cc index 7151b16..be394ba 100644 --- a/scribo/sandbox/icdar_13_table/src/main.cc +++ b/scribo/sandbox/icdar_13_table/src/main.cc @@ -8,6 +8,7 @@ #include <mln/fun/v2v/rgb_to_luma.hh> #include <mln/io/pbm/all.hh> +#include <mln/io/pdf/load.hh> #include <mln/io/ppm/all.hh> #include <mln/labeling/all.hh> @@ -180,6 +181,10 @@ void find_borders(image2d<bool>& ima, bottom = find_bottom(ima); } + /********/ + /* MAIN */ + /********/ + int main(int argc, char** argv) { typedef value::label_16 V; @@ -195,116 +200,139 @@ int main(int argc, char** argv) V nhlines, nvlines; L hlines_ima, vlines_ima; scribo::component_set<L> hlines, vlines; + std::ostringstream path; - // PARAMETERS + // Parameters unsigned av_height = 4; unsigned av_width = 4; int delta_prox_h = 5; unsigned min_height = 71; unsigned min_width = 31; + unsigned dpi = 72; - // Loadin and binarization + // Loading and binarization std::ofstream xml; start_xml(xml, "final.xml", argv[1]); - io::ppm::load(original, argv[1]); - filtered = data::transform(original, mln::fun::v2v::rgb_to_luma<value::int_u8>()); + //io::ppm::load(original, argv[1]); + util::array< image2d<value::rgb8> > pdf; + io::pdf::load(pdf, argv[1], dpi); + for (unsigned page = 0; page < pdf.nelements(); ++page) + { + original = pdf[page]; + filtered = data::transform(original, mln::fun::v2v::rgb_to_luma<value::int_u8>()); - bin = scribo::binarization::sauvola(filtered, 81, 0.44); - final = data::convert(value::rgb8(), bin); + bin = scribo::binarization::sauvola(filtered, 81, 0.44); + final = data::convert(value::rgb8(), bin); - initialize(mask, bin); - initialize(ima_texts, bin); - initialize(ima_tables, bin); - data::fill(ima_tables, false); + initialize(mask, bin); + initialize(ima_texts, bin); + initialize(ima_tables, bin); + data::fill(ima_tables, false); - bin_without_lines = duplicate(bin); + bin_without_lines = duplicate(bin); - // Lines extraction - hlines = scribo::primitive::extract::lines_h_discontinued(bin, c4(), nhlines, min_width, 2); - vlines = scribo::primitive::extract::lines_v_discontinued(bin, c4(), nvlines, min_height, 2); + // Lines extraction + hlines = scribo::primitive::extract::lines_h_discontinued(bin, c4(), nhlines, min_width, 2); + vlines = scribo::primitive::extract::lines_v_discontinued(bin, c4(), nvlines, min_height, 2); - get_horizontal_lines(hlines, ima_tables, bin_without_lines, av_height); - get_vertical_lines(vlines, ima_tables, bin_without_lines, av_width, delta_prox_h); + get_horizontal_lines(hlines, ima_tables, bin_without_lines, av_height); + get_vertical_lines(vlines, ima_tables, bin_without_lines, av_width, delta_prox_h); - // Denoising - bin_without_lines_denoised = scribo::preprocessing::denoise_fg(bin_without_lines, c8(), 3); + // Denoising + bin_without_lines_denoised = scribo::preprocessing::denoise_fg(bin_without_lines, c8(), 4); - // Set the tables mask - unsigned n; - labeled = labeling::blobs(ima_tables, c8(), n); - masks = scribo::component_set< image2d<unsigned> >(labeled, n); - data::fill(mask, false); + // Set the tables mask + unsigned n; + labeled = labeling::blobs(ima_tables, c8(), n); + masks = scribo::component_set< image2d<unsigned> >(labeled, n); + data::fill(mask, false); - for (unsigned i = 1; i <= masks.nelements(); ++i) - data::fill((mask | masks(i).bbox()).rw(), true); + for (unsigned i = 1; i <= masks.nelements(); ++i) + data::fill((mask | masks(i).bbox()).rw(), true); - // Compose table zones with bin_without_lines_lines - ima_texts = logical::and_(bin_without_lines_denoised, mask); + // Compose table zones with bin_without_lines_lines + ima_texts = logical::and_(bin_without_lines_denoised, mask); - // Isolate texts between tables - for (unsigned i = 1; i <= masks.nelements(); ++i) - { - image2d<bool> table_mask, isolated_text; - std::ostringstream path; - bool empty = true; + // Isolate texts between tables + for (unsigned i = 1; i <= masks.nelements(); ++i) + { + image2d<bool> table_mask, isolated_text; + bool empty = true; - initialize(table_mask, bin); - data::fill(table_mask, false); - data::fill((table_mask | masks(i).bbox()).rw(), true); + initialize(table_mask, bin); + data::fill(table_mask, false); + data::fill((table_mask | masks(i).bbox()).rw(), true); - isolated_text = logical::and_(bin_without_lines_denoised, table_mask); + isolated_text = logical::and_(bin_without_lines_denoised, table_mask); - mln_piter_(image2d<bool>) p(isolated_text.domain()); + mln_piter_(image2d<bool>) p(isolated_text.domain()); - for_all(p) - empty = empty && !(isolated_text(p)); + for_all(p) + empty = empty && !(isolated_text(p)); - if (!empty) - { - path << "output/8_" << i << "_isolated.pbm"; - io::pbm::save(isolated_text, path.str()); + if (!empty) + { + path.str(""); + path << "output/p" << page << "_8_" << i << "_isolated.pbm"; + io::pbm::save(isolated_text, path.str()); - // Find coordinated - unsigned left, right, top, bottom; - find_borders(isolated_text, left, right, top, bottom); + // Find coordinates + unsigned left, right, top, bottom; + find_borders(isolated_text, left, right, top, bottom); - std::cout << "(" << left << "," << top << ") ->" - << "(" << right << "," << bottom << ")" << std::endl; + point2d p1, p2, p3, p4; - point2d p1, p2, p3, p4; + p1 = point2d(top, left); + p2 = point2d(top, right); + p3 = point2d(bottom, right); + p4 = point2d(bottom, left); - p1 = point2d(top, left); - p2 = point2d(top, right); - p3 = point2d(bottom, right); - p4 = point2d(bottom, left); + draw::line(final, p1, p2, literal::green); + draw::line(final, p2, p3, literal::green); + draw::line(final, p3, p4, literal::green); + draw::line(final, p4, p1, literal::green); - draw::line(final, p1, p2, literal::red); - draw::line(final, p2, p3, literal::red); - draw::line(final, p3, p4, literal::red); - draw::line(final, p4, p1, literal::red); - - write_table(xml, p1, p3); + write_table(xml, p1, p3); + } } - } - // Get lines images - hlines_ima = hlines.labeled_image(); - vlines_ima = vlines.labeled_image(); - ima_hlines = data::convert(bool(), hlines_ima); - ima_vlines = data::convert(bool(), vlines_ima); - - // Write images and close XML - io::pbm::save(bin, "output/0_bin.pbm"); - io::pbm::save(bin_without_lines, "output/1_bin_without_lines.pbm"); - io::pbm::save(bin_without_lines_denoised, "output/2_bin_without_lines_denoised.pbm"); - io::pbm::save(ima_hlines, "output/3_hlines.pbm"); - io::pbm::save(ima_vlines, "output/4_vlines.pbm"); - io::pbm::save(ima_tables, "output/5_tables.pbm"); - io::pbm::save(mask, "output/6_mask.pbm"); - io::pbm::save(ima_texts, "output/7_texts.pbm"); - /* Save 8_i_isolated */ - io::ppm::save(final, "output/9_final.ppm"); + // Get lines images + hlines_ima = hlines.labeled_image(); + vlines_ima = vlines.labeled_image(); + ima_hlines = data::convert(bool(), hlines_ima); + ima_vlines = data::convert(bool(), vlines_ima); + + // Write images and close XML + path.str(""); path << "output/p" << page << "_0_bin.pbm"; + io::pbm::save(bin, path.str()); + + path.str(""); path << "output/p" << page << "_1_bin_without_lines.pbm"; + io::pbm::save(bin_without_lines, path.str()); + + path.str(""); path << "output/p" << page << "_2_bin_without_lines_denoised.pbm"; + io::pbm::save(bin_without_lines_denoised, path.str()); + + path.str(""); path << "output/p" << page << "_3_hlines.pbm"; + io::pbm::save(ima_hlines, path.str()); + + path.str(""); path << "output/p" << page << "_4_vlines.pbm"; + io::pbm::save(ima_vlines, path.str()); + + path.str(""); path << "output/p" << page << "_5_tables.pbm"; + io::pbm::save(ima_tables, path.str()); + + path.str(""); path << "output/p" << page << "_6_mask.pbm"; + io::pbm::save(mask, path.str()); + + path.str(""); path << "output/p" << page << "_7_texts.pbm"; + io::pbm::save(ima_texts, path.str()); + + /* Save 8_i_isolated */ + + path.str(""); path << "output/p" << page << "_9_final.pbm"; + io::ppm::save(final, path.str()); + } end_xml(xml); -- 1.7.2.5

1 0

branch sandbox/anthony updated: olena-2.0-387-g99b3631
by git＠lrde.epita.fr 11 Mar '13

11 Mar '13

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 sandbox/anthony has been updated via 99b36312a31ce549eb18103576794165e61281f3 (commit) from 023ba4982dddb83f3bde1d78d849f025a8f70d8c (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 ----------------------------------------------------------------- 99b3631 Output a simple and valid XML and image of bounding boxes + Add TODO file ----------------------------------------------------------------------- Summary of changes: scribo/sandbox/icdar_13_table/TODO | 12 ++++++++++++ scribo/sandbox/icdar_13_table/src/main.cc | 25 ++++++++++++++++++++++--- 2 files changed, 34 insertions(+), 3 deletions(-) create mode 100644 scribo/sandbox/icdar_13_table/TODO hooks/post-receive -- Olena, a generic and efficient image processing platform

1 0

olena-2.0-387-g99b3631 Output a simple and valid XML and image of bounding boxes + Add TODO file
by Anthony Seure 11 Mar '13

11 Mar '13

--- scribo/sandbox/icdar_13_table/TODO | 12 ++++++++++++ scribo/sandbox/icdar_13_table/src/main.cc | 25 ++++++++++++++++++++++--- 2 files changed, 34 insertions(+), 3 deletions(-) create mode 100644 scribo/sandbox/icdar_13_table/TODO diff --git a/scribo/sandbox/icdar_13_table/TODO b/scribo/sandbox/icdar_13_table/TODO new file mode 100644 index 0000000..b34e7a2 --- /dev/null +++ b/scribo/sandbox/icdar_13_table/TODO @@ -0,0 +1,12 @@ +#------------------------------------------------------------------------------# + TODO LIST - ICDAR 2013 - Table competition +#------------------------------------------------------------------------------# + +Table location sub-competition : + * Load PDF files (instead of PNM) + * Find links betwwen pages for mutlipages tables + * Get text from reversed-video zones + * Expand the process to borderless tables + +Table structure recognition sub-competition : + * All diff --git a/scribo/sandbox/icdar_13_table/src/main.cc b/scribo/sandbox/icdar_13_table/src/main.cc index e4dbb11..7151b16 100644 --- a/scribo/sandbox/icdar_13_table/src/main.cc +++ b/scribo/sandbox/icdar_13_table/src/main.cc @@ -1,7 +1,9 @@ #include <mln/binarization/all.hh> #include <mln/core/image/image2d.hh> + #include <mln/data/all.hh> +#include <mln/draw/line.hh> #include <mln/fun/v2v/rgb_to_luma.hh> @@ -9,6 +11,7 @@ #include <mln/io/ppm/all.hh> #include <mln/labeling/all.hh> +#include <mln/literal/all.hh> #include <mln/logical/and.hh> #include <mln/value/all.hh> @@ -45,7 +48,8 @@ void write_table(std::ofstream& xml, const point2d& start, const point2d& end) << "\t\t<region id='" << region << "' page='" << page << "'>" << std::endl << "\t\t<bounding-box x1='" << start[1] << "' y1='" << start[0] << "' " << "x2='" << end[1] << "' y2='" << end[0] << "'/>" << std::endl - << "\t\t</region>" << std::endl; + << "\t\t</region>" << std::endl + << "\t</table>" << std::endl; ++table; } @@ -181,7 +185,7 @@ int main(int argc, char** argv) typedef value::label_16 V; typedef image2d<V> L; - image2d<value::rgb8> original; + image2d<value::rgb8> original, final; image2d<value::int_u8> filtered; image2d<bool> bin, ima_hlines, ima_vlines, ima_tables, ima_texts, bin_without_lines, bin_without_lines_denoised, @@ -207,6 +211,7 @@ int main(int argc, char** argv) filtered = data::transform(original, mln::fun::v2v::rgb_to_luma<value::int_u8>()); bin = scribo::binarization::sauvola(filtered, 81, 0.44); + final = data::convert(value::rgb8(), bin); initialize(mask, bin); initialize(ima_texts, bin); @@ -266,9 +271,22 @@ int main(int argc, char** argv) std::cout << "(" << left << "," << top << ") ->" << "(" << right << "," << bottom << ")" << std::endl; + + point2d p1, p2, p3, p4; + + p1 = point2d(top, left); + p2 = point2d(top, right); + p3 = point2d(bottom, right); + p4 = point2d(bottom, left); + + draw::line(final, p1, p2, literal::red); + draw::line(final, p2, p3, literal::red); + draw::line(final, p3, p4, literal::red); + draw::line(final, p4, p1, literal::red); + + write_table(xml, p1, p3); } } - // Get lines images hlines_ima = hlines.labeled_image(); @@ -286,6 +304,7 @@ int main(int argc, char** argv) io::pbm::save(mask, "output/6_mask.pbm"); io::pbm::save(ima_texts, "output/7_texts.pbm"); /* Save 8_i_isolated */ + io::ppm::save(final, "output/9_final.ppm"); end_xml(xml); -- 1.7.2.5

1 0

branch sandbox/anthony updated: olena-2.0-386-g023ba49
by git＠lrde.epita.fr 11 Mar '13

11 Mar '13

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 sandbox/anthony has been updated via 023ba4982dddb83f3bde1d78d849f025a8f70d8c (commit) from 28f05e67c5ca54c7fbaf2b37a9630ecc746fa78f (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 ----------------------------------------------------------------- 023ba49 Find bounding coordinates for each tables ----------------------------------------------------------------------- Summary of changes: scribo/sandbox/icdar_13_table/src/main.cc | 122 +++++++++++++++++++++++++++++ 1 files changed, 122 insertions(+), 0 deletions(-) hooks/post-receive -- Olena, a generic and efficient image processing platform

1 0

olena-2.0-386-g023ba49 Find bounding coordinates for each tables
by Anthony Seure 11 Mar '13

11 Mar '13

--- scribo/sandbox/icdar_13_table/src/main.cc | 122 +++++++++++++++++++++++++++++ 1 files changed, 122 insertions(+), 0 deletions(-) diff --git a/scribo/sandbox/icdar_13_table/src/main.cc b/scribo/sandbox/icdar_13_table/src/main.cc index 2ed125d..e4dbb11 100644 --- a/scribo/sandbox/icdar_13_table/src/main.cc +++ b/scribo/sandbox/icdar_13_table/src/main.cc @@ -88,6 +88,94 @@ void get_vertical_lines(scribo::component_set<L>& vlines, } } +unsigned find_left(image2d<bool>& ima) +{ + bool found = false; + + def::coord col = geom::min_col(ima); + while (col < geom::max_col(ima) && !found) + { + def::coord row = geom::min_row(ima); + while (row < geom::max_row(ima) && !found) + { + found = opt::at(ima, row, col); + ++row; + } + ++col; + } + + return (col - 1); +} + +unsigned find_right(image2d<bool>& ima) +{ + bool found = false; + + def::coord col = geom::max_col(ima) - 1; + while (col >= geom::min_col(ima) && !found) + { + def::coord row = geom::min_row(ima); + while (row < geom::max_row(ima) && !found) + { + found = opt::at(ima, row, col); + ++row; + } + --col; + } + + return (col + 1); + +} + +unsigned find_top(image2d<bool>& ima) +{ + bool found = false; + + def::coord row = geom::min_row(ima); + while (row < geom::max_row(ima) && !found) + { + def::coord col = geom::min_col(ima); + while (col < geom::max_col(ima) && !found) + { + found = opt::at(ima, row, col); + ++col; + } + ++row; + } + + return (row - 1); + +} + +unsigned find_bottom(image2d<bool>& ima) +{ + bool found = false; + + def::coord row = geom::max_row(ima) - 1; + while (row >= geom::min_row(ima) && !found) + { + def::coord col = geom::min_col(ima); + while (col < geom::max_col(ima) && !found) + { + found = opt::at(ima, row, col); + ++col; + } + --row; + } + + return (row + 1); +} + +void find_borders(image2d<bool>& ima, + unsigned& left, unsigned& right, + unsigned& top, unsigned& bottom) +{ + left = find_left(ima); + right = find_right(ima); + top = find_top(ima); + bottom = find_bottom(ima); +} + int main(int argc, char** argv) { typedef value::label_16 V; @@ -149,6 +237,39 @@ int main(int argc, char** argv) // Compose table zones with bin_without_lines_lines ima_texts = logical::and_(bin_without_lines_denoised, mask); + // Isolate texts between tables + for (unsigned i = 1; i <= masks.nelements(); ++i) + { + image2d<bool> table_mask, isolated_text; + std::ostringstream path; + bool empty = true; + + initialize(table_mask, bin); + data::fill(table_mask, false); + data::fill((table_mask | masks(i).bbox()).rw(), true); + + isolated_text = logical::and_(bin_without_lines_denoised, table_mask); + + mln_piter_(image2d<bool>) p(isolated_text.domain()); + + for_all(p) + empty = empty && !(isolated_text(p)); + + if (!empty) + { + path << "output/8_" << i << "_isolated.pbm"; + io::pbm::save(isolated_text, path.str()); + + // Find coordinated + unsigned left, right, top, bottom; + find_borders(isolated_text, left, right, top, bottom); + + std::cout << "(" << left << "," << top << ") ->" + << "(" << right << "," << bottom << ")" << std::endl; + } + } + + // Get lines images hlines_ima = hlines.labeled_image(); vlines_ima = vlines.labeled_image(); @@ -164,6 +285,7 @@ int main(int argc, char** argv) io::pbm::save(ima_tables, "output/5_tables.pbm"); io::pbm::save(mask, "output/6_mask.pbm"); io::pbm::save(ima_texts, "output/7_texts.pbm"); + /* Save 8_i_isolated */ end_xml(xml); -- 1.7.2.5

1 0

branch sandbox/anthony created: olena-2.0-385-g28f05e6
by git＠lrde.epita.fr 08 Mar '13

08 Mar '13

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 sandbox/anthony has been created at 28f05e67c5ca54c7fbaf2b37a9630ecc746fa78f (commit) - Log ----------------------------------------------------------------- 28f05e6 [ICDAR_13] Find texts inside cells on cropped tables ----------------------------------------------------------------------- hooks/post-receive -- Olena, a generic and efficient image processing platform

1 0