March 2011 - Olena-patches

last-svn-commit-793-ge93e375 Small fixes in viewer.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

* demo/viewer/main.cc: Add more includes. * demo/viewer/viewer.cc: Add debug output. --- scribo/ChangeLog | 8 ++++++++ scribo/demo/viewer/main.cc | 5 ++++- scribo/demo/viewer/viewer.cc | 8 +++++++- 3 files changed, 19 insertions(+), 2 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index 9e63c2a..358201c 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,3 +1,11 @@ +2011-03-11 Guillaume Lazzara <z(a)lrde.epita.fr> + + Small fixes in viewer. + + * demo/viewer/main.cc: Add more includes. + + * demo/viewer/viewer.cc: Add debug output. + 2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> * scribo/io/xml/internal/extended_page_xml_visitor.hh: Save diff --git a/scribo/demo/viewer/main.cc b/scribo/demo/viewer/main.cc index 6d1cb18..069d5e9 100644 --- a/scribo/demo/viewer/main.cc +++ b/scribo/demo/viewer/main.cc @@ -1,4 +1,5 @@ -// Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) +// Copyright (C) 2010, 2011 EPITA Research and Development Laboratory +// (LRDE) // // This file is part of Olena. // @@ -23,6 +24,8 @@ #include <mln/math/pi.hh> #include <mln/io/magick/load.hh> #include <scribo/make/debug_filename.hh> +#include <mln/debug/filename.hh> +#include <mln/literal/white.hh> int main(int argc, char** argv) { diff --git a/scribo/demo/viewer/viewer.cc b/scribo/demo/viewer/viewer.cc index f34b3e7..173cb3b 100644 --- a/scribo/demo/viewer/viewer.cc +++ b/scribo/demo/viewer/viewer.cc @@ -372,7 +372,10 @@ Viewer::add_text(QDomNode line) coords = coords.nextSibling(); if (coords.isNull()) + { + qDebug() << "Warning : textline without coordinates..."; return; + } QDomNode point = coords.firstChild(); @@ -393,7 +396,7 @@ Viewer::add_text(QDomNode line) point = point.nextSibling(); } - QString text = line.toElement().attribute("text", "none"); + QString text = line.toElement().attribute("text", ""); QFont font("Times"); font.setPixelSize(a_height + d_height); QGraphicsTextItem* text_item = scene_->addText(text, font); @@ -417,7 +420,10 @@ Viewer::add_region(QDomNode father, QString attr_id) coords = coords.nextSibling(); if (coords.isNull()) + { + qDebug() << "Warning : add_region - region without coordinates"; return; + } QDomNode point = coords.firstChild(); QVector<QPoint> points; -- 1.5.6.5

1 0

last-svn-commit-792-g446d792 scribo/io/xml/internal/extended_page_xml_visitor.hh: Save separators data.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

--- scribo/ChangeLog | 5 +++++ .../io/xml/internal/extended_page_xml_visitor.hh | 4 ++++ 2 files changed, 9 insertions(+), 0 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index 3f17098..9e63c2a 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,5 +1,10 @@ 2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> + * scribo/io/xml/internal/extended_page_xml_visitor.hh: Save + separators data. + +2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> + * scribo/io/xml/internal/full_xml_visitor.hh: Do not copy image data. diff --git a/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh b/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh index 4da0755..f573d88 100644 --- a/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh +++ b/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh @@ -113,6 +113,10 @@ namespace scribo if (doc.has_elements()) doc.elements().accept(*this); + // line seraparators + if (doc.has_line_seps()) + doc.line_seps_comps().accept(*this); + // Whitespace seraparators if (doc.has_whitespace_seps()) doc.whitespace_seps_comps().accept(*this); -- 1.5.6.5

1 0

last-svn-commit-791-g79532c9 scribo/io/xml/internal/full_xml_visitor.hh: Do not copy image data.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

--- scribo/ChangeLog | 5 +++++ scribo/scribo/io/xml/internal/full_xml_visitor.hh | 11 ++++++----- 2 files changed, 11 insertions(+), 5 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index 2af84c0..3f17098 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,5 +1,10 @@ 2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> + * scribo/io/xml/internal/full_xml_visitor.hh: Do not copy image + data. + +2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> + Make OCR step optional in viewer. * demo/viewer/Makefile.am: Add new files. diff --git a/scribo/scribo/io/xml/internal/full_xml_visitor.hh b/scribo/scribo/io/xml/internal/full_xml_visitor.hh index 2b09b3a..8576c27 100644 --- a/scribo/scribo/io/xml/internal/full_xml_visitor.hh +++ b/scribo/scribo/io/xml/internal/full_xml_visitor.hh @@ -279,8 +279,9 @@ namespace scribo // FIXME: Try to avoid that! border::resize(lbl, 0); - QByteArray lbl64((const char *)lbl.buffer(), - lbl.nelements() * sizeof(mln_value(L))); + QByteArray + lbl64 = QByteArray::fromRawData((const char *)lbl.buffer(), + lbl.nelements() * sizeof(mln_value(L))); lbl64 = qCompress(lbl64, COMPRESSION_LEVEL); lbl64 = lbl64.toBase64(); @@ -299,8 +300,9 @@ namespace scribo << "<![CDATA["; border::resize(seps, 0); - QByteArray seps64((const char *)seps.buffer(), - seps.nelements() * sizeof(bool)); + QByteArray + seps64 = QByteArray::fromRawData((const char *)seps.buffer(), + seps.nelements() * sizeof(bool)); seps64 = qCompress(seps64, COMPRESSION_LEVEL); seps64 = seps64.toBase64(); @@ -346,7 +348,6 @@ namespace scribo break; } - default: case component::Image: { -- 1.5.6.5

1 0

last-svn-commit-790-g3c3cd93 Make OCR step optional in viewer.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

* demo/viewer/Makefile.am: Add new files. * demo/viewer/config.cc, * demo/viewer/config.hh: Add OCR settings. * demo/viewer/toolchain_options.cc, * demo/viewer/toolchain_options.hh, * demo/viewer/toolchain_options.ui: Remove. * demo/viewer/ocr_options.cc, * demo/viewer/ocr_options.hh, * demo/viewer/ocr_options.ui, * demo/viewer/preprocessing_options.cc, * demo/viewer/preprocessing_options.hh, * demo/viewer/preprocessing_options.ui, * demo/viewer/segmentation_options.cc, * demo/viewer/segmentation_options.hh, * demo/viewer/segmentation_options.ui: New option widgets. * demo/viewer/preferences_dialog.cc: Insert new widgets. * demo/viewer/runner.cc: Take OCR settings into account. --- scribo/ChangeLog | 30 +++++++- scribo/demo/viewer/Makefile.am | 20 ++++- scribo/demo/viewer/config.cc | 28 ++++++- scribo/demo/viewer/config.hh | 11 ++- scribo/demo/viewer/ocr_options.cc | 74 ++++++++++++++++ scribo/demo/viewer/ocr_options.hh | 40 +++++++++ scribo/demo/viewer/ocr_options.ui | 62 ++++++++++++++ scribo/demo/viewer/preferences_dialog.cc | 26 +++++- scribo/demo/viewer/preprocessing_options.cc | 66 +++++++++++++++ scribo/demo/viewer/preprocessing_options.hh | 38 +++++++++ scribo/demo/viewer/preprocessing_options.ui | 76 +++++++++++++++++ scribo/demo/viewer/runner.cc | 3 +- scribo/demo/viewer/segmentation_options.cc | 56 ++++++++++++ scribo/demo/viewer/segmentation_options.hh | 37 ++++++++ scribo/demo/viewer/segmentation_options.ui | 52 ++++++++++++ scribo/demo/viewer/toolchain_options.cc | 75 ----------------- scribo/demo/viewer/toolchain_options.hh | 37 -------- scribo/demo/viewer/toolchain_options.ui | 121 --------------------------- 18 files changed, 605 insertions(+), 247 deletions(-) create mode 100644 scribo/demo/viewer/ocr_options.cc create mode 100644 scribo/demo/viewer/ocr_options.hh create mode 100644 scribo/demo/viewer/ocr_options.ui create mode 100644 scribo/demo/viewer/preprocessing_options.cc create mode 100644 scribo/demo/viewer/preprocessing_options.hh create mode 100644 scribo/demo/viewer/preprocessing_options.ui create mode 100644 scribo/demo/viewer/segmentation_options.cc create mode 100644 scribo/demo/viewer/segmentation_options.hh create mode 100644 scribo/demo/viewer/segmentation_options.ui delete mode 100644 scribo/demo/viewer/toolchain_options.cc delete mode 100644 scribo/demo/viewer/toolchain_options.hh delete mode 100644 scribo/demo/viewer/toolchain_options.ui diff --git a/scribo/ChangeLog b/scribo/ChangeLog index f4a56fd..2af84c0 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,6 +1,34 @@ 2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> - * scribo/toolchain/internal/content_in_doc_functor.hh: Make OCR step optional. + Make OCR step optional in viewer. + + * demo/viewer/Makefile.am: Add new files. + + * demo/viewer/config.cc, + * demo/viewer/config.hh: Add OCR settings. + + * demo/viewer/toolchain_options.cc, + * demo/viewer/toolchain_options.hh, + * demo/viewer/toolchain_options.ui: Remove. + + * demo/viewer/ocr_options.cc, + * demo/viewer/ocr_options.hh, + * demo/viewer/ocr_options.ui, + * demo/viewer/preprocessing_options.cc, + * demo/viewer/preprocessing_options.hh, + * demo/viewer/preprocessing_options.ui, + * demo/viewer/segmentation_options.cc, + * demo/viewer/segmentation_options.hh, + * demo/viewer/segmentation_options.ui: New option widgets. + + * demo/viewer/preferences_dialog.cc: Insert new widgets. + + * demo/viewer/runner.cc: Take OCR settings into account. + +2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> + + * scribo/toolchain/internal/content_in_doc_functor.hh: Make OCR + step optional. 2011-03-03 Guillaume Lazzara <z(a)lrde.epita.fr> diff --git a/scribo/demo/viewer/Makefile.am b/scribo/demo/viewer/Makefile.am index 2c5e4cf..09d9ecc 100644 --- a/scribo/demo/viewer/Makefile.am +++ b/scribo/demo/viewer/Makefile.am @@ -30,7 +30,9 @@ scribo_viewer_SOURCES = viewer.cc \ image_region.cc \ help_dialog.cc \ step_widget.cc \ - toolchain_options.cc \ + preprocessing_options.cc\ + segmentation_options.cc \ + ocr_options.cc \ runner.cc \ preprocess.cc \ process.cc \ @@ -65,8 +67,12 @@ BUILT_SOURCES = viewer.moc.cc \ image_region.moc.cc \ help_dialog.ui.h \ step_widget.moc.cc \ - toolchain_options.moc.cc \ - toolchain_options.ui.h \ + preprocessing_options.moc.cc \ + preprocessing_options.ui.h \ + segmentation_options.moc.cc \ + segmentation_options.ui.h \ + ocr_options.moc.cc \ + ocr_options.ui.h \ runner.moc.cc \ preprocess.moc.cc \ process.moc.cc \ @@ -91,7 +97,9 @@ noinst_HEADERS = viewer.hh \ common.hh \ help_dialog.hh \ step_widget.hh \ - toolchain_options.hh \ + preprocessing_options.hh\ + segmentation_options.hh\ + ocr_options.hh \ runner.hh \ preprocess.hh \ process.hh \ @@ -121,7 +129,9 @@ EXTRA_DIST = \ general_options.ui \ preferences_dialog.ui \ README \ - toolchain_options.ui \ + preprocessing_options.ui\ + segmentation_options.ui \ + ocr_options.ui \ help_dialog.ui \ demodir.hh.in diff --git a/scribo/demo/viewer/config.cc b/scribo/demo/viewer/config.cc index bf12e95..6bdcd35 100644 --- a/scribo/demo/viewer/config.cc +++ b/scribo/demo/viewer/config.cc @@ -1,4 +1,5 @@ -// Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) +// Copyright (C) 2010, 2011 EPITA Research and Development Laboratory +// (LRDE) // // This file is part of Olena. // @@ -108,6 +109,31 @@ void config::set_segmentation_find_seps(int seps) } +// OCR + +bool config::ocr_enabled() +{ + return value("ocr/enabled", true).toBool(); +} + +void config::set_ocr_enabled(bool b) +{ + setValue("ocr/enabled", b); +} + + + +QString config::ocr_language() +{ + return value("ocr/language", 0).toString(); +} + +void config::set_ocr_language(const QString& lang) +{ + setValue("ocr/language", lang); +} + + // General options diff --git a/scribo/demo/viewer/config.hh b/scribo/demo/viewer/config.hh index 07d193c..74dab3b 100644 --- a/scribo/demo/viewer/config.hh +++ b/scribo/demo/viewer/config.hh @@ -1,4 +1,5 @@ -// Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) +// Copyright (C) 2010, 2011 EPITA Research and Development Laboratory +// (LRDE) // // This file is part of Olena. // @@ -49,6 +50,14 @@ public: void set_segmentation_find_seps(int seps); + // OCR + bool ocr_enabled(); + void set_ocr_enabled(bool b); + + QString ocr_language(); + void set_ocr_language(const QString& lang); + + // General options bool general_save_xml_enabled(); void set_general_save_xml_enabled(bool b); diff --git a/scribo/demo/viewer/ocr_options.cc b/scribo/demo/viewer/ocr_options.cc new file mode 100644 index 0000000..0ff09d9 --- /dev/null +++ b/scribo/demo/viewer/ocr_options.cc @@ -0,0 +1,74 @@ +// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) +// +// This file is part of Olena. +// +// Olena is free software: you can redistribute it and/or modify it under +// the terms of the GNU General Public License as published by the Free +// Software Foundation, version 2 of the License. +// +// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. + +#include "defs.hh" +#include "ocr_options.hh" +#include "config.hh" + +// Defines enum of binarization algorithms +# include <scribo/toolchain/internal/text_in_doc_preprocess_functor.hh> + +using namespace scribo::toolchain::internal; + + +static const char *language[][2] = { + { "English", "eng" }, + { "French", "fra" }, + { 0, 0 } +}; + + +ocr_options::ocr_options(QWidget *parent) + : OptionWidget(parent) +{ + setupUi(this); + + for (unsigned i = 0; language[i][0]; ++i) + ocr_language->insertItem(i, language[i][0]); + + load_config(); +} + +ocr_options::~ocr_options() +{ +} + + +int ocr_options::find_index(const QString& lang) +{ + for (unsigned i = 0; language[i][0]; ++i) + if (lang == language[i][1]) + return i; + return 0; +} + + +void ocr_options::load_config() +{ + config * const conf = config::get_instance(); + + enable_ocr->setChecked(conf->ocr_enabled()); + ocr_language->setCurrentIndex(find_index(conf->ocr_language())); +} + + +void ocr_options::save_config() +{ + config * const conf = config::get_instance(); + + conf->set_ocr_enabled(enable_ocr->isChecked()); + conf->set_ocr_language(language[ocr_language->currentIndex()][1]); +} diff --git a/scribo/demo/viewer/ocr_options.hh b/scribo/demo/viewer/ocr_options.hh new file mode 100644 index 0000000..67ab4ef --- /dev/null +++ b/scribo/demo/viewer/ocr_options.hh @@ -0,0 +1,40 @@ +// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) +// +// This file is part of Olena. +// +// Olena is free software: you can redistribute it and/or modify it under +// the terms of the GNU General Public License as published by the Free +// Software Foundation, version 2 of the License. +// +// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. + +#ifndef SCRIBO_DEMO_VIEWER_OCR_OPTIONS_HH +# define SCRIBO_DEMO_VIEWER_OCR_OPTIONS_HH + +# include <QtGui> +# include <ocr_options.ui.h> +# include "option_widget.hh" + +class ocr_options : public OptionWidget, private Ui::OcrOptions +{ + Q_OBJECT; + +public: + ocr_options(QWidget *parent = 0); + ~ocr_options(); + + void load_config(); + void save_config(); + +private: + int find_index(const QString& lang); + +}; + +#endif // ! SCRIBO_DEMO_VIEWER_OCR_OPTIONS_HH diff --git a/scribo/demo/viewer/ocr_options.ui b/scribo/demo/viewer/ocr_options.ui new file mode 100644 index 0000000..d001c64 --- /dev/null +++ b/scribo/demo/viewer/ocr_options.ui @@ -0,0 +1,62 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>OcrOptions</class> + <widget class="QWidget" name="OcrOptions"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>366</width> + <height>112</height> + </rect> + </property> + <property name="windowTitle"> + <string>Form</string> + </property> + <layout class="QVBoxLayout" name="verticalLayout"> + <item> + <widget class="QGroupBox" name="enable_ocr"> + <property name="title"> + <string>Enable OCR</string> + </property> + <property name="alignment"> + <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignVCenter</set> + </property> + <property name="flat"> + <bool>true</bool> + </property> + <property name="checkable"> + <bool>true</bool> + </property> + <layout class="QGridLayout" name="gridLayout"> + <item row="1" column="1"> + <spacer name="verticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>40</height> + </size> + </property> + </spacer> + </item> + <item row="0" column="0"> + <widget class="QLabel" name="label"> + <property name="text"> + <string>Language</string> + </property> + </widget> + </item> + <item row="0" column="1"> + <widget class="QComboBox" name="ocr_language"/> + </item> + </layout> + </widget> + </item> + </layout> + </widget> + <resources/> + <connections/> +</ui> diff --git a/scribo/demo/viewer/preferences_dialog.cc b/scribo/demo/viewer/preferences_dialog.cc index dfd6061..1eb80a7 100644 --- a/scribo/demo/viewer/preferences_dialog.cc +++ b/scribo/demo/viewer/preferences_dialog.cc @@ -1,4 +1,5 @@ -// Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) +// Copyright (C) 2010, 2011 EPITA Research and Development Laboratory +// (LRDE) // // This file is part of Olena. // @@ -15,8 +16,10 @@ // along with Olena. If not, see <http://www.gnu.org/licenses/>. #include "preferences_dialog.hh" -#include "toolchain_options.hh" #include "general_options.hh" +#include "ocr_options.hh" +#include "preprocessing_options.hh" +#include "segmentation_options.hh" @@ -41,7 +44,7 @@ preferences_dialog::~preferences_dialog() void preferences_dialog::load_option_list() { - static const char *options[] = { "General", "Toolchain", 0 }; + static const char *options[] = { "General", "Preprocessing", "Segmentation", "OCR", 0 }; int i; for (i = 0; options[i]; ++i) @@ -66,19 +69,32 @@ void preferences_dialog::select_option_widget(int row) if (!widgets_[row]) { + QWidget *widget = 0; + switch (row) { case 0: - widgets_[0] = new general_options(this); + widget = new general_options(this); break; case 1: - widgets_[1] = new toolchain_options(this); + widget = new preprocessing_options(this); + break; + + case 2: + widget = new segmentation_options(this); + break; + + case 3: + widget = new ocr_options(this); break; default: qDebug() << "select_option_widget - Hu? Something wrong..."; } + + if (widget) + widgets_[row] = widget; } if (horizontalLayout_2->count() == 2) diff --git a/scribo/demo/viewer/preprocessing_options.cc b/scribo/demo/viewer/preprocessing_options.cc new file mode 100644 index 0000000..2d55fea --- /dev/null +++ b/scribo/demo/viewer/preprocessing_options.cc @@ -0,0 +1,66 @@ +// Copyright (C) 2010, 2011 EPITA Research and Development Laboratory +// (LRDE) +// +// This file is part of Olena. +// +// Olena is free software: you can redistribute it and/or modify it under +// the terms of the GNU General Public License as published by the Free +// Software Foundation, version 2 of the License. +// +// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. + +#include "defs.hh" +#include "preprocessing_options.hh" +#include "config.hh" + +// Defines enum of binarization algorithms +# include <scribo/toolchain/internal/text_in_doc_preprocess_functor.hh> + +using namespace scribo::toolchain::internal; + +preprocessing_options::preprocessing_options(QWidget *parent) + : OptionWidget(parent) +{ + setupUi(this); + + bin_algoCbox->insertItem(Convert, "Violent convert"); + bin_algoCbox->insertItem(Sauvola, "Local threshold"); + bin_algoCbox->insertItem(SauvolaMs, + "Local threshold multiscale"); + + load_config(); +} + +preprocessing_options::~preprocessing_options() +{ +} + + +void preprocessing_options::load_config() +{ + config * const conf = config::get_instance(); + + subsampleCb->setChecked(conf->preprocessing_subsample()); + remove_bgCb->setChecked(conf->preprocessing_remove_bg()); + deskewCb->setChecked(conf->preprocessing_deskew()); + remove_noiseCb->setChecked(conf->preprocessing_remove_noise()); + bin_algoCbox->setCurrentIndex(conf->preprocessing_bin_algo()); +} + + +void preprocessing_options::save_config() +{ + config * const conf = config::get_instance(); + + conf->set_preprocessing_subsample(subsampleCb->isChecked()); + conf->set_preprocessing_remove_bg(remove_bgCb->isChecked()); + conf->set_preprocessing_deskew(deskewCb->isChecked()); + conf->set_preprocessing_remove_noise(remove_noiseCb->isChecked()); + conf->set_preprocessing_bin_algo(bin_algoCbox->currentIndex()); +} diff --git a/scribo/demo/viewer/preprocessing_options.hh b/scribo/demo/viewer/preprocessing_options.hh new file mode 100644 index 0000000..c676880 --- /dev/null +++ b/scribo/demo/viewer/preprocessing_options.hh @@ -0,0 +1,38 @@ +// Copyright (C) 2010, 2011 EPITA Research and Development Laboratory +// (LRDE) +// +// This file is part of Olena. +// +// Olena is free software: you can redistribute it and/or modify it under +// the terms of the GNU General Public License as published by the Free +// Software Foundation, version 2 of the License. +// +// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. + +#ifndef SCRIBO_DEMO_VIEWER_PREPROCESSING_OPTIONS_HH +# define SCRIBO_DEMO_VIEWER_PREPROCESSING_OPTIONS_HH + +# include <QtGui> +# include <preprocessing_options.ui.h> +# include "option_widget.hh" + +class preprocessing_options : public OptionWidget, private Ui::PreprocessingOptions +{ + Q_OBJECT; + +public: + preprocessing_options(QWidget *parent = 0); + ~preprocessing_options(); + + void load_config(); + void save_config(); + +}; + +#endif // ! SCRIBO_DEMO_VIEWER_PREPROCESSING_OPTIONS_HH diff --git a/scribo/demo/viewer/preprocessing_options.ui b/scribo/demo/viewer/preprocessing_options.ui new file mode 100644 index 0000000..962cf49 --- /dev/null +++ b/scribo/demo/viewer/preprocessing_options.ui @@ -0,0 +1,76 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>PreprocessingOptions</class> + <widget class="QWidget" name="PreprocessingOptions"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>466</width> + <height>198</height> + </rect> + </property> + <property name="windowTitle"> + <string>Configure toolchain</string> + </property> + <layout class="QGridLayout" name="gridLayout"> + <item row="0" column="0"> + <layout class="QHBoxLayout" name="horizontalLayout_3"> + <item> + <widget class="QLabel" name="label_3"> + <property name="text"> + <string>Binarization method:</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="bin_algoCbox"/> + </item> + </layout> + </item> + <item row="1" column="0"> + <widget class="QCheckBox" name="subsampleCb"> + <property name="text"> + <string>Run on subsampled image</string> + </property> + </widget> + </item> + <item row="2" column="0"> + <widget class="QCheckBox" name="remove_bgCb"> + <property name="text"> + <string>Remove background (slow)</string> + </property> + </widget> + </item> + <item row="3" column="0"> + <widget class="QCheckBox" name="deskewCb"> + <property name="text"> + <string>Deskew</string> + </property> + </widget> + </item> + <item row="4" column="0"> + <widget class="QCheckBox" name="remove_noiseCb"> + <property name="text"> + <string>Remove noise</string> + </property> + </widget> + </item> + <item row="5" column="0"> + <spacer name="verticalSpacer"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>48</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + <resources/> + <connections/> +</ui> diff --git a/scribo/demo/viewer/runner.cc b/scribo/demo/viewer/runner.cc index a3cc883..e923db7 100644 --- a/scribo/demo/viewer/runner.cc +++ b/scribo/demo/viewer/runner.cc @@ -156,7 +156,8 @@ void runner::process(const image2d<value::rgb8>& original_ima, || find_seps == defs::LinesAndWhitespaces); f.enable_whitespace_seps = (find_seps == defs::Whitespaces || find_seps == defs::LinesAndWhitespaces); - + f.enable_ocr = conf->ocr_enabled(); + f.ocr_language = conf->ocr_language().toAscii().data(); f.xml_format = scribo::io::xml::PageExtended; diff --git a/scribo/demo/viewer/segmentation_options.cc b/scribo/demo/viewer/segmentation_options.cc new file mode 100644 index 0000000..c120879 --- /dev/null +++ b/scribo/demo/viewer/segmentation_options.cc @@ -0,0 +1,56 @@ +// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) +// +// This file is part of Olena. +// +// Olena is free software: you can redistribute it and/or modify it under +// the terms of the GNU General Public License as published by the Free +// Software Foundation, version 2 of the License. +// +// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. + +#include "defs.hh" +#include "segmentation_options.hh" +#include "config.hh" + +// Defines enum of binarization algorithms +# include <scribo/toolchain/internal/text_in_doc_preprocess_functor.hh> + +using namespace scribo::toolchain::internal; + +segmentation_options::segmentation_options(QWidget *parent) + : OptionWidget(parent) +{ + setupUi(this); + + find_sepsCbox->insertItem(defs::Lines, "Lines"); + find_sepsCbox->insertItem(defs::Whitespaces, "Whitespaces"); + find_sepsCbox->insertItem(defs::LinesAndWhitespaces, "Lines and whitespaces"); + + load_config(); +} + +segmentation_options::~segmentation_options() +{ +} + + +void segmentation_options::load_config() +{ + config * const conf = config::get_instance(); + + find_sepsCbox->setCurrentIndex(conf->segmentation_find_seps()); +} + + +void segmentation_options::save_config() +{ + config * const conf = config::get_instance(); + + conf->set_segmentation_find_seps(find_sepsCbox->currentIndex()); +} diff --git a/scribo/demo/viewer/segmentation_options.hh b/scribo/demo/viewer/segmentation_options.hh new file mode 100644 index 0000000..bdf65f9 --- /dev/null +++ b/scribo/demo/viewer/segmentation_options.hh @@ -0,0 +1,37 @@ +// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) +// +// This file is part of Olena. +// +// Olena is free software: you can redistribute it and/or modify it under +// the terms of the GNU General Public License as published by the Free +// Software Foundation, version 2 of the License. +// +// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. + +#ifndef SCRIBO_DEMO_VIEWER_SEGMENTATION_OPTIONS_HH +# define SCRIBO_DEMO_VIEWER_SEGMENTATION_OPTIONS_HH + +# include <QtGui> +# include <segmentation_options.ui.h> +# include "option_widget.hh" + +class segmentation_options : public OptionWidget, private Ui::SegmentationOptions +{ + Q_OBJECT; + +public: + segmentation_options(QWidget *parent = 0); + ~segmentation_options(); + + void load_config(); + void save_config(); + +}; + +#endif // ! SCRIBO_DEMO_VIEWER_SEGMENTATION_OPTIONS_HH diff --git a/scribo/demo/viewer/segmentation_options.ui b/scribo/demo/viewer/segmentation_options.ui new file mode 100644 index 0000000..485e448 --- /dev/null +++ b/scribo/demo/viewer/segmentation_options.ui @@ -0,0 +1,52 @@ +<?xml version="1.0" encoding="UTF-8"?> +<ui version="4.0"> + <class>SegmentationOptions</class> + <widget class="QWidget" name="SegmentationOptions"> + <property name="geometry"> + <rect> + <x>0</x> + <y>0</y> + <width>400</width> + <height>300</height> + </rect> + </property> + <property name="windowTitle"> + <string>Form</string> + </property> + <layout class="QGridLayout" name="gridLayout"> + <item row="0" column="0"> + <layout class="QHBoxLayout" name="horizontalLayout"> + <item> + <widget class="QLabel" name="label"> + <property name="text"> + <string>Find separators</string> + </property> + </widget> + </item> + <item> + <widget class="QComboBox" name="find_sepsCbox"> + <property name="editable"> + <bool>false</bool> + </property> + </widget> + </item> + </layout> + </item> + <item row="1" column="0"> + <spacer name="verticalSpacer_2"> + <property name="orientation"> + <enum>Qt::Vertical</enum> + </property> + <property name="sizeHint" stdset="0"> + <size> + <width>20</width> + <height>258</height> + </size> + </property> + </spacer> + </item> + </layout> + </widget> + <resources/> + <connections/> +</ui> diff --git a/scribo/demo/viewer/toolchain_options.cc b/scribo/demo/viewer/toolchain_options.cc deleted file mode 100644 index 6024ff3..0000000 --- a/scribo/demo/viewer/toolchain_options.cc +++ /dev/null @@ -1,75 +0,0 @@ -// Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) -// -// This file is part of Olena. -// -// Olena is free software: you can redistribute it and/or modify it under -// the terms of the GNU General Public License as published by the Free -// Software Foundation, version 2 of the License. -// -// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. - -#include "defs.hh" -#include "toolchain_options.hh" -#include "config.hh" - -// Defines enum of binarization algorithms -# include <scribo/toolchain/internal/text_in_doc_preprocess_functor.hh> - -using namespace scribo::toolchain::internal; - -toolchain_options::toolchain_options(QWidget *parent) - : OptionWidget(parent) -{ - setupUi(this); - - bin_algoCbox->insertItem(Convert, "Violent convert"); - bin_algoCbox->insertItem(Sauvola, "Local threshold"); - bin_algoCbox->insertItem(SauvolaMs, - "Local threshold multiscale"); - - find_sepsCbox->insertItem(defs::Lines, "Lines"); - find_sepsCbox->insertItem(defs::Whitespaces, "Whitespaces"); - find_sepsCbox->insertItem(defs::LinesAndWhitespaces, "Lines and whitespaces"); - - load_config(); -} - -toolchain_options::~toolchain_options() -{ -} - - -void toolchain_options::load_config() -{ - config * const conf = config::get_instance(); - - // Preprocessing - subsampleCb->setChecked(conf->preprocessing_subsample()); - remove_bgCb->setChecked(conf->preprocessing_remove_bg()); - deskewCb->setChecked(conf->preprocessing_deskew()); - remove_noiseCb->setChecked(conf->preprocessing_remove_noise()); - bin_algoCbox->setCurrentIndex(conf->preprocessing_bin_algo()); - - // Page segmentation - find_sepsCbox->setCurrentIndex(conf->segmentation_find_seps()); -} - - -void toolchain_options::save_config() -{ - config * const conf = config::get_instance(); - - conf->set_preprocessing_subsample(subsampleCb->isChecked()); - conf->set_preprocessing_remove_bg(remove_bgCb->isChecked()); - conf->set_preprocessing_deskew(deskewCb->isChecked()); - conf->set_preprocessing_remove_noise(remove_noiseCb->isChecked()); - conf->set_preprocessing_bin_algo(bin_algoCbox->currentIndex()); - - conf->set_segmentation_find_seps(find_sepsCbox->currentIndex()); -} diff --git a/scribo/demo/viewer/toolchain_options.hh b/scribo/demo/viewer/toolchain_options.hh deleted file mode 100644 index afad3e7..0000000 --- a/scribo/demo/viewer/toolchain_options.hh +++ /dev/null @@ -1,37 +0,0 @@ -// Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) -// -// This file is part of Olena. -// -// Olena is free software: you can redistribute it and/or modify it under -// the terms of the GNU General Public License as published by the Free -// Software Foundation, version 2 of the License. -// -// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. - -#ifndef SCRIBO_DEMO_VIEWER_TOOLCHAIN_OPTIONS_HH -# define SCRIBO_DEMO_VIEWER_TOOLCHAIN_OPTIONS_HH - -# include <QtGui> -# include <toolchain_options.ui.h> -# include "option_widget.hh" - -class toolchain_options : public OptionWidget, private Ui::ToolchainOptions -{ - Q_OBJECT; - -public: - toolchain_options(QWidget *parent = 0); - ~toolchain_options(); - - void load_config(); - void save_config(); - -}; - -#endif // ! SCRIBO_DEMO_VIEWER_TOOLCHAIN_OPTIONS_HH diff --git a/scribo/demo/viewer/toolchain_options.ui b/scribo/demo/viewer/toolchain_options.ui deleted file mode 100644 index 748696d..0000000 --- a/scribo/demo/viewer/toolchain_options.ui +++ /dev/null @@ -1,121 +0,0 @@ -<?xml version="1.0" encoding="UTF-8"?> -<ui version="4.0"> - <class>ToolchainOptions</class> - <widget class="QWidget" name="ToolchainOptions"> - <property name="geometry"> - <rect> - <x>0</x> - <y>0</y> - <width>492</width> - <height>264</height> - </rect> - </property> - <property name="windowTitle"> - <string>Configure toolchain</string> - </property> - <layout class="QVBoxLayout" name="verticalLayout_3"> - <item> - <widget class="QGroupBox" name="groupBox"> - <property name="title"> - <string>Preprocessing</string> - </property> - <layout class="QVBoxLayout" name="verticalLayout"> - <item> - <widget class="QCheckBox" name="subsampleCb"> - <property name="text"> - <string>Run on subsampled image</string> - </property> - </widget> - </item> - <item> - <widget class="QCheckBox" name="remove_bgCb"> - <property name="text"> - <string>Remove background (slow)</string> - </property> - </widget> - </item> - <item> - <widget class="QCheckBox" name="deskewCb"> - <property name="text"> - <string>Deskew</string> - </property> - </widget> - </item> - <item> - <widget class="QCheckBox" name="remove_noiseCb"> - <property name="text"> - <string>Remove noise</string> - </property> - </widget> - </item> - <item> - <layout class="QHBoxLayout" name="horizontalLayout_3"> - <item> - <widget class="QLabel" name="label_3"> - <property name="text"> - <string>Binarization method:</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="bin_algoCbox"/> - </item> - </layout> - </item> - <item> - <spacer name="verticalSpacer"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - </layout> - </widget> - </item> - <item> - <widget class="QGroupBox" name="groupBox_2"> - <property name="title"> - <string>Page segmentation</string> - </property> - <layout class="QVBoxLayout" name="verticalLayout_2"> - <item> - <layout class="QHBoxLayout" name="horizontalLayout"> - <item> - <widget class="QLabel" name="label"> - <property name="text"> - <string>Find separators</string> - </property> - </widget> - </item> - <item> - <widget class="QComboBox" name="find_sepsCbox"/> - </item> - </layout> - </item> - <item> - <spacer name="verticalSpacer_2"> - <property name="orientation"> - <enum>Qt::Vertical</enum> - </property> - <property name="sizeHint" stdset="0"> - <size> - <width>20</width> - <height>40</height> - </size> - </property> - </spacer> - </item> - </layout> - </widget> - </item> - </layout> - </widget> - <resources/> - <connections/> -</ui> -- 1.5.6.5

1 0

last-svn-commit-789-g8461438 scribo/toolchain/internal/content_in_doc_functor.hh: Make OCR step optional.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

--- scribo/ChangeLog | 8 ++++++-- .../toolchain/internal/content_in_doc_functor.hh | 19 ++++++++++++------- 2 files changed, 18 insertions(+), 9 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index 52da198..f4a56fd 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,9 +1,13 @@ +2011-03-04 Guillaume Lazzara <z(a)lrde.epita.fr> + + * scribo/toolchain/internal/content_in_doc_functor.hh: Make OCR step optional. + 2011-03-03 Guillaume Lazzara <z(a)lrde.epita.fr> Compress image data in XML output. - * scribo/scribo/io/xml/internal/full_xml_visitor.hh, - * scribo/scribo/io/xml/load.hh: Here. + * scribo/io/xml/internal/full_xml_visitor.hh, + * scribo/io/xml/load.hh: Here. 2011-03-03 Guillaume Lazzara <z(a)lrde.epita.fr> diff --git a/scribo/scribo/toolchain/internal/content_in_doc_functor.hh b/scribo/scribo/toolchain/internal/content_in_doc_functor.hh index dcbb4f7..40f013c 100644 --- a/scribo/scribo/toolchain/internal/content_in_doc_functor.hh +++ b/scribo/scribo/toolchain/internal/content_in_doc_functor.hh @@ -111,6 +111,7 @@ namespace scribo bool enable_denoising; bool enable_line_seps; bool enable_whitespace_seps; + bool enable_ocr; bool enable_debug; bool save_doc_as_xml; scribo::io::xml::Format xml_format; @@ -136,6 +137,7 @@ namespace scribo : enable_denoising(true), enable_line_seps(true), enable_whitespace_seps(true), + enable_ocr(true), enable_debug(false), save_doc_as_xml(false), xml_format(scribo::io::xml::PageExtended), @@ -373,11 +375,14 @@ namespace scribo // Text recognition - on_new_progress_label("Recognizing text"); - scribo::text::recognition(lines, ocr_language.c_str()); + if (enable_ocr) + { + on_new_progress_label("Recognizing text"); - on_progress(); + scribo::text::recognition(lines, ocr_language.c_str()); + on_progress(); + } // Link text lines on_new_progress_label("Linking text lines"); @@ -390,7 +395,7 @@ namespace scribo image2d<value::rgb8> debug = data::convert(value::rgb8(), original_image); for_all_lines(l, lines) { - if (! lines(l).is_valid() || lines(l).is_hidden() || lines(l).type() != line::Text) + if (! lines(l).is_textline()) continue; mln::draw::box(debug, lines(l).bbox(), literal::blue); @@ -427,7 +432,7 @@ namespace scribo mln::util::array<accu::shape::bbox<point2d> > nbbox(llinks.nelements()); for_all_lines(i, lines) { - if (! lines(i).is_valid() || lines(i).is_hidden() || lines(i).type() != line::Text) + if (! lines(i).is_textline()) continue; mln::draw::box(debug, lines(i).bbox(), literal::red); @@ -497,8 +502,8 @@ namespace scribo int content_in_doc_functor<I>::nsteps() const { - return 11 + enable_denoising + enable_line_seps - + enable_whitespace_seps + save_doc_as_xml; + return 10 + enable_denoising + enable_line_seps + + enable_whitespace_seps + enable_ocr + save_doc_as_xml; } -- 1.5.6.5

1 0

last-svn-commit-788-ge334996 Compress image data in XML output.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

* scribo/scribo/io/xml/internal/full_xml_visitor.hh, * scribo/scribo/io/xml/load.hh: Here. --- scribo/ChangeLog | 7 ++++ scribo/scribo/io/xml/internal/full_xml_visitor.hh | 34 ++++++++++++++------ scribo/scribo/io/xml/load.hh | 12 +++++-- 3 files changed, 39 insertions(+), 14 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index da0d095..52da198 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,5 +1,12 @@ 2011-03-03 Guillaume Lazzara <z(a)lrde.epita.fr> + Compress image data in XML output. + + * scribo/scribo/io/xml/internal/full_xml_visitor.hh, + * scribo/scribo/io/xml/load.hh: Here. + +2011-03-03 Guillaume Lazzara <z(a)lrde.epita.fr> + Small fixes in Scribo. * scribo/convert/from_base64.hh: Remove guards HAVE_QT. diff --git a/scribo/scribo/io/xml/internal/full_xml_visitor.hh b/scribo/scribo/io/xml/internal/full_xml_visitor.hh index 9c5bd1d..2b09b3a 100644 --- a/scribo/scribo/io/xml/internal/full_xml_visitor.hh +++ b/scribo/scribo/io/xml/internal/full_xml_visitor.hh @@ -31,6 +31,9 @@ /// XML serializer Visitor. # include <fstream> + +# include <QByteArray> + # include <scribo/core/internal/doc_xml_serializer.hh> # include <scribo/core/document.hh> # include <scribo/core/component_set.hh> @@ -40,11 +43,12 @@ # include <scribo/core/line_links.hh> # include <scribo/core/line_info.hh> -# include <scribo/convert/to_base64.hh> - # include <scribo/io/xml/internal/print_box_coords.hh> # include <scribo/io/xml/internal/print_page_preambule.hh> +// Compression level 0-9. 9 is the best but is slow. +// 5 seems to be a good compromise. +# define COMPRESSION_LEVEL 5 namespace scribo { @@ -273,15 +277,20 @@ namespace scribo << "\" width=\"" << lbl.domain().width() << "\">" << "<![CDATA["; - util::array<unsigned char> lbl64; - convert::to_base64(lbl, lbl64); - output.write((const char *)lbl64.std_vector().data(), - lbl64.nelements()); + // FIXME: Try to avoid that! + border::resize(lbl, 0); + QByteArray lbl64((const char *)lbl.buffer(), + lbl.nelements() * sizeof(mln_value(L))); + lbl64 = qCompress(lbl64, COMPRESSION_LEVEL); + lbl64 = lbl64.toBase64(); + + output.write(lbl64.data(), lbl64.size()); output << "]]></labeled_image>" << std::endl; } // Save separators image + if (comp_set.has_separators()) { const mln_ch_value(L,bool)& seps = comp_set.separators(); output << "<separators_image " @@ -289,10 +298,13 @@ namespace scribo << "\" width=\"" << seps.domain().width() << "\">" << "<![CDATA["; - util::array<unsigned char> seps64; - convert::to_base64(seps, seps64); - output.write((const char *)seps64.std_vector().data(), - seps64.nelements()); + border::resize(seps, 0); + QByteArray seps64((const char *)seps.buffer(), + seps.nelements() * sizeof(bool)); + seps64 = qCompress(seps64, COMPRESSION_LEVEL); + seps64 = seps64.toBase64(); + + output.write(seps64.data(), seps64.size()); output << "]]></separators_image>" << std::endl; } @@ -453,4 +465,6 @@ namespace scribo } // end of namespace scribo +# undef COMPRESSION_LEVEL + #endif // SCRIBO_IO_XML_INTERNAL_FULL_XML_VISITOR_HH diff --git a/scribo/scribo/io/xml/load.hh b/scribo/scribo/io/xml/load.hh index e0f4548..8042c75 100644 --- a/scribo/scribo/io/xml/load.hh +++ b/scribo/scribo/io/xml/load.hh @@ -50,8 +50,6 @@ # include <scribo/core/line_set.hh> # include <scribo/core/line_info.hh> -# include <scribo/convert/from_base64.hh> - namespace scribo { @@ -405,13 +403,19 @@ namespace scribo { case LabeledImage: { - scribo::convert::from_base64(ch, comp_set_data->ima_); + QByteArray data = ch.toAscii(); + data = QByteArray::fromBase64(data); + data = qUncompress(data); + memcpy((char *) comp_set_data->ima_.buffer(), data.data(), data.size()); } break; case SeparatorsImage: { - scribo::convert::from_base64(ch, comp_set_data->separators_); + QByteArray data = ch.toAscii(); + data = QByteArray::fromBase64(data); + data = qUncompress(data); + memcpy((char *) comp_set_data->separators_.buffer(), data.data(), data.size()); } break; -- 1.5.6.5

1 0

last-svn-commit-787-g8ef4bff Small fixes in Scribo.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

* scribo/convert/from_base64.hh: Remove guards HAVE_QT. * scribo/core/document.hh: Initialize ImageMagick. * scribo/io/xml/internal/html_markups_replace.hh: Remove. Move code... * scribo/core/line_info.hh: ... here. * scribo/core/line_links.hh: Make use of is_textline method. * scribo/core/paragraph_info.hh: Add nlines method. * scribo/core/paragraph_set.hh: Add a construction routine using a line_set. * scribo/io/xml/internal/extended_page_xml_visitor.hh, * scribo/io/xml/internal/print_box_coords.hh, * scribo/io/xml/internal/print_page_preambule.hh, * src/primitive/extract/separators_nonvisible.cc: Fix includes. * scribo/io/xml/save.hh: Add a precondition. * scribo/text/extract_lines.hh, * src/debug/show_info_median_inter_characters.cc, * src/debug/show_info_x_height.cc, * src/debug/show_text_lines.cc: Fix label type. --- scribo/ChangeLog | 30 ++++++ scribo/scribo/convert/from_base64.hh | 8 -- scribo/scribo/core/document.hh | 1 + scribo/scribo/core/line_info.hh | 45 ++++++++- scribo/scribo/core/line_links.hh | 5 +- scribo/scribo/core/paragraph_info.hh | 9 ++ scribo/scribo/core/paragraph_set.hh | 28 ++++++ .../io/xml/internal/extended_page_xml_visitor.hh | 2 - .../scribo/io/xml/internal/html_markups_replace.hh | 97 -------------------- scribo/scribo/io/xml/internal/print_box_coords.hh | 3 + .../scribo/io/xml/internal/print_page_preambule.hh | 5 + scribo/scribo/io/xml/save.hh | 2 + scribo/scribo/text/extract_lines.hh | 16 ++-- .../src/debug/show_info_median_inter_characters.cc | 12 +- scribo/src/debug/show_info_x_height.cc | 12 +- scribo/src/debug/show_text_lines.cc | 2 +- .../src/primitive/extract/separators_nonvisible.cc | 2 +- 17 files changed, 143 insertions(+), 136 deletions(-) delete mode 100644 scribo/scribo/io/xml/internal/html_markups_replace.hh diff --git a/scribo/ChangeLog b/scribo/ChangeLog index a300b37..da0d095 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,3 +1,33 @@ +2011-03-03 Guillaume Lazzara <z(a)lrde.epita.fr> + + Small fixes in Scribo. + + * scribo/convert/from_base64.hh: Remove guards HAVE_QT. + + * scribo/core/document.hh: Initialize ImageMagick. + + * scribo/io/xml/internal/html_markups_replace.hh: Remove. Move code... + * scribo/core/line_info.hh: ... here. + + * scribo/core/line_links.hh: Make use of is_textline method. + + * scribo/core/paragraph_info.hh: Add nlines method. + + * scribo/core/paragraph_set.hh: Add a construction routine using a + line_set. + + * scribo/io/xml/internal/extended_page_xml_visitor.hh, + * scribo/io/xml/internal/print_box_coords.hh, + * scribo/io/xml/internal/print_page_preambule.hh, + * src/primitive/extract/separators_nonvisible.cc: Fix includes. + + * scribo/io/xml/save.hh: Add a precondition. + + * scribo/text/extract_lines.hh, + * src/debug/show_info_median_inter_characters.cc, + * src/debug/show_info_x_height.cc, + * src/debug/show_text_lines.cc: Fix label type. + 2011-03-02 Guillaume Lazzara <z(a)lrde.epita.fr> Regen mk files. diff --git a/scribo/scribo/convert/from_base64.hh b/scribo/scribo/convert/from_base64.hh index 8c14a0f..5eeeb53 100644 --- a/scribo/scribo/convert/from_base64.hh +++ b/scribo/scribo/convert/from_base64.hh @@ -82,13 +82,9 @@ namespace scribo Image<I>& output); -# if defined HAVE_QT - template <typename I> void from_base64(const QString& data64, Image<I>& output_); -# endif // ! HAVE_QT - # ifndef MLN_INCLUDE_ONLY @@ -191,8 +187,6 @@ namespace scribo } -# if defined HAVE_QT - template <typename I> void from_base64(const QString& data64, Image<I>& output_) @@ -205,8 +199,6 @@ namespace scribo trace::exiting("scribo::convert::to_base64"); } -# endif // ! HAVE_QT - # endif // ! MLN_INCLUDE_ONLY diff --git a/scribo/scribo/core/document.hh b/scribo/scribo/core/document.hh index 372f0a4..689d0e3 100644 --- a/scribo/scribo/core/document.hh +++ b/scribo/scribo/core/document.hh @@ -160,6 +160,7 @@ namespace scribo void document<L>::open() { + Magick::InitializeMagick(0); mln::io::magick::load(image_, filename_); } diff --git a/scribo/scribo/core/line_info.hh b/scribo/scribo/core/line_info.hh index 33a1529..46ba720 100644 --- a/scribo/scribo/core/line_info.hh +++ b/scribo/scribo/core/line_info.hh @@ -35,9 +35,10 @@ /// from the x_height and the baseline. /// /// \fixme The way the meanline and the baseline are computed is not -/// optimal and does not work is the components are too high (because +/// optimal and does not work if the components are too high (because /// of the median accumulator and int_u12 overflows). +# include <map> # include <mln/core/alias/box2d.hh> # include <mln/core/alias/point2d.hh> @@ -53,8 +54,6 @@ # include <scribo/core/line_set.hh> # include <scribo/core/component_set.hh> -# include <scribo/io/xml/internal/html_markups_replace.hh> - # include <scribo/core/concept/serializable.hh> @@ -287,6 +286,42 @@ namespace scribo namespace internal { + // INTERNAL TOOLS + + static inline std::map<char, std::string> init_map() + { + std::map<char, std::string> html_map; + html_map['\"'] = """; + html_map['<'] = "<"; + html_map['>'] = ">"; + html_map['&'] = "&"; + return html_map; + } + + + inline + std::string + html_markups_replace(const std::string& input) + { + static std::map<char, std::string> map = init_map(); + + std::string output = input; + for (unsigned i = 0; i < input.size(); ++i) + { + std::map<char, std::string>::iterator it = map.find(output.at(i)); + if (it != map.end()) + { + output.replace(i, 1, it->second); + i += it->second.size() - 1; + } + } + return output; + } + + + + // LINE INFO DATA + template <typename L> line_info_data<L>::line_info_data() { @@ -641,7 +676,7 @@ namespace scribo line_info<L>::update_text(const std::string& str) { data_->text_ = str; - data_->html_text_ = scribo::io::xml::internal::html_markups_replace(str); + data_->html_text_ = scribo::internal::html_markups_replace(str); } @@ -869,6 +904,8 @@ namespace scribo ref_line = mln::math::min(comp_set(c).bbox().pmin().row(), ref_line); } + // FIXME: compute font color! + for_all_elements(i, data_->components_) { unsigned c = data_->components_(i); diff --git a/scribo/scribo/core/line_links.hh b/scribo/scribo/core/line_links.hh index fdd09a5..ab36a73 100644 --- a/scribo/scribo/core/line_links.hh +++ b/scribo/scribo/core/line_links.hh @@ -215,9 +215,8 @@ namespace scribo line_links<L>::init() { for (unsigned i = 0; i < nelements(); ++i) - if (! data_->lines_(i).is_valid() - || data_->lines_(i).is_hidden() - || data_->lines_(i).type() != line::Text) + if (!data_->lines_(i).is_valid() + || !data_->lines_(i).is_textline()) { data_->line_to_link_(i) = 0; } diff --git a/scribo/scribo/core/paragraph_info.hh b/scribo/scribo/core/paragraph_info.hh index 17f847f..f4b1309 100644 --- a/scribo/scribo/core/paragraph_info.hh +++ b/scribo/scribo/core/paragraph_info.hh @@ -52,6 +52,8 @@ namespace scribo const mln::util::array<line_id_t>& line_ids() const; + unsigned nlines() const; + bool is_valid() const; private: @@ -106,6 +108,13 @@ namespace scribo } template <typename L> + unsigned + paragraph_info<L>::nlines() const + { + return line_ids_.nelements(); + } + + template <typename L> bool paragraph_info<L>::is_valid() const { diff --git a/scribo/scribo/core/paragraph_set.hh b/scribo/scribo/core/paragraph_set.hh index 5451069..f081472 100644 --- a/scribo/scribo/core/paragraph_set.hh +++ b/scribo/scribo/core/paragraph_set.hh @@ -215,6 +215,7 @@ namespace scribo } // end of namespace scribo::make::internal + // FIXME: move that code into paragraph_set constructor? template <typename L> scribo::paragraph_set<L> paragraph(const line_links<L>& llinks) @@ -241,6 +242,33 @@ namespace scribo return parset; } + + // FIXME: move that code into paragraph_set constructor? + template <typename L> + scribo::paragraph_set<L> + paragraph(const scribo::line_set<L>& lines) + { + line_links<L> links(lines); + links.init(); + + unsigned npars; + mln::fun::i2v::array<unsigned> + par_ids = mln::make::relabelfun(links.line_to_link(), + links.nelements() - 1, npars); + paragraph_set<L> parset(links, npars); + + for_all_links(l, links) + if (links(l)) + { + value::int_u16 par_id = par_ids(l); + parset(par_id).add_line(lines(l)); + } + + return parset; + } + + + } // end of namespace scribo::make diff --git a/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh b/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh index 5d8a672..4da0755 100644 --- a/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh +++ b/scribo/scribo/io/xml/internal/extended_page_xml_visitor.hh @@ -40,8 +40,6 @@ # include <scribo/core/line_links.hh> # include <scribo/core/line_info.hh> -# include <scribo/convert/to_base64.hh> - # include <scribo/io/xml/internal/print_box_coords.hh> # include <scribo/io/xml/internal/print_page_preambule.hh> diff --git a/scribo/scribo/io/xml/internal/html_markups_replace.hh b/scribo/scribo/io/xml/internal/html_markups_replace.hh deleted file mode 100644 index 76f8107..0000000 --- a/scribo/scribo/io/xml/internal/html_markups_replace.hh +++ /dev/null @@ -1,97 +0,0 @@ -// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) -// -// This file is part of Olena. -// -// Olena is free software: you can redistribute it and/or modify it under -// the terms of the GNU General Public License as published by the Free -// Software Foundation, version 2 of the License. -// -// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. -// -// As a special exception, you may use this file as part of a free -// software project 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 SCRIBO_IO_XML_INTERNAL_HTML_MARKUPS_REPLACE_HH -# define SCRIBO_IO_XML_INTERNAL_HTML_MARKUPS_REPLACE_HH - -/// \file -/// -/// \brief Replace HTML markups characters by their corresponding -/// markups. - - -namespace scribo -{ - - namespace io - { - - namespace xml - { - - namespace internal - { - - /*! \brief Replace HTML markups characters by their corresponding - markups. - */ - inline - std::string - html_markups_replace(std::string& input); - - -# ifndef MLN_INCLUDE_ONLY - - static inline std::map<char, std::string> init_map() - { - std::map<char, std::string> html_map; - html_map['\"'] = """; - html_map['<'] = "<"; - html_map['>'] = ">"; - html_map['&'] = "&"; - return html_map; - } - - - inline - std::string - html_markups_replace(const std::string& input) - { - static std::map<char, std::string> map = init_map(); - - std::string output = input; - for (unsigned i = 0; i < input.size(); ++i) - { - std::map<char, std::string>::iterator it = map.find(output.at(i)); - if (it != map.end()) - { - output.replace(i, 1, it->second); - i += it->second.size() - 1; - } - } - return output; - } - -# endif // ! MLN_INCLUDE_ONLY - - } // end of namespace scribo::io::xml::internal - - } // end of namespace scribo::io::xml - - } // end of namespace scribo::io - -} // end of namespace scribo - -#endif // ! SCRIBO_IO_XML_INTERNAL_HTML_MARKUPS_REPLACE_HH diff --git a/scribo/scribo/io/xml/internal/print_box_coords.hh b/scribo/scribo/io/xml/internal/print_box_coords.hh index d3aeedf..ad84709 100644 --- a/scribo/scribo/io/xml/internal/print_box_coords.hh +++ b/scribo/scribo/io/xml/internal/print_box_coords.hh @@ -30,6 +30,7 @@ /// /// \brief Prints box2d coordinates to XML data. +# include <fstream> # include <mln/core/alias/box2d.hh> namespace scribo @@ -44,6 +45,8 @@ namespace scribo namespace internal { + using namespace mln; + /*! \brief Prints box2d coordinates to XML data. */ void diff --git a/scribo/scribo/io/xml/internal/print_page_preambule.hh b/scribo/scribo/io/xml/internal/print_page_preambule.hh index b5ae891..3ee29be 100644 --- a/scribo/scribo/io/xml/internal/print_page_preambule.hh +++ b/scribo/scribo/io/xml/internal/print_page_preambule.hh @@ -30,7 +30,9 @@ /// /// \brief Print PAGE XML format preambule. +# include <fstream> # include <mln/core/alias/box2d.hh> +# include <scribo/core/document.hh> namespace scribo { @@ -44,6 +46,9 @@ namespace scribo namespace internal { + using namespace mln; + + /// \brief Print PAGE XML format preambule. template <typename L> void print_PAGE_preambule(std::ofstream& output, diff --git a/scribo/scribo/io/xml/save.hh b/scribo/scribo/io/xml/save.hh index 30579d0..7428e5d 100644 --- a/scribo/scribo/io/xml/save.hh +++ b/scribo/scribo/io/xml/save.hh @@ -121,6 +121,8 @@ namespace scribo { trace::entering("scribo::io::xml::save"); + mln_precondition(doc.is_open()); + // Open file std::ofstream output(output_name.c_str()); if (! output) diff --git a/scribo/scribo/text/extract_lines.hh b/scribo/scribo/text/extract_lines.hh index b81cb79..9949d09 100644 --- a/scribo/scribo/text/extract_lines.hh +++ b/scribo/scribo/text/extract_lines.hh @@ -33,8 +33,8 @@ # include <mln/core/concept/image.hh> -# include <mln/value/int_u16.hh> +# include <scribo/core/def/lbl_type.hh> # include <scribo/core/line_set.hh> # include <scribo/primitive/extract/components.hh> @@ -71,13 +71,13 @@ namespace scribo ** \return A set of lines. */ template <typename I, typename N> - line_set<mln_ch_value(I,value::int_u16)> + line_set<mln_ch_value(I,scribo::def::lbl_type)> extract_lines(const Image<I>& input_, const Neighborhood<N>& nbh_, const mln_ch_value(I,bool)& separators); /// \overload template <typename I, typename N> - line_set<mln_ch_value(I,value::int_u16)> + line_set<mln_ch_value(I,scribo::def::lbl_type)> extract_lines(const Image<I>& input, const Neighborhood<N>& nbh); @@ -85,7 +85,7 @@ namespace scribo template <typename I, typename N> - line_set<mln_ch_value(I,value::int_u16)> + line_set<mln_ch_value(I,scribo::def::lbl_type)> extract_lines(const Image<I>& input, const Neighborhood<N>& nbh) { mln_ch_value(I,bool) seps; @@ -94,7 +94,7 @@ namespace scribo template <typename I, typename N> - line_set<mln_ch_value(I,value::int_u16)> + line_set<mln_ch_value(I,scribo::def::lbl_type)> extract_lines(const Image<I>& input_, const Neighborhood<N>& nbh_, const mln_ch_value(I,bool)& separators) { @@ -107,10 +107,10 @@ namespace scribo mln_precondition(nbh.is_valid()); /// Finding comps. - typedef mln_ch_value(I,value::int_u16) L; - value::int_u16 ncomps; + typedef mln_ch_value(I,scribo::def::lbl_type) L; + scribo::def::lbl_type ncomps; component_set<L> - comps = scribo::primitive::extract::components(input, c8(), ncomps); + comps = scribo::primitive::extract::components(input, nbh, ncomps); /// First filtering. comps = scribo::filter::components_small(comps, 3); diff --git a/scribo/src/debug/show_info_median_inter_characters.cc b/scribo/src/debug/show_info_median_inter_characters.cc index b55079d..2ba240e 100644 --- a/scribo/src/debug/show_info_median_inter_characters.cc +++ b/scribo/src/debug/show_info_median_inter_characters.cc @@ -1,5 +1,5 @@ -// Copyright (C) 2009, 2010 EPITA Research and Development Laboratory -// (LRDE) +// Copyright (C) 2009, 2010, 2011 EPITA Research and Development +// Laboratory (LRDE) // // This file is part of Olena. // @@ -35,7 +35,7 @@ #include <mln/io/ppm/save.hh> #include <mln/value/rgb8.hh> -#include <mln/value/label_16.hh> +#include <mln/value/int_u16.hh> #include <mln/draw/box.hh> #include <mln/draw/line.hh> @@ -71,14 +71,14 @@ int main(int argc, char* argv[]) image2d<bool> input; - io::pbm::load(input, argv[1]); + mln::io::pbm::load(input, argv[1]); - typedef image2d<value::label_16> L; + typedef image2d<scribo::def::lbl_type> L; line_set<L> line = scribo::text::extract_lines(input, c8()); image2d<value::rgb8> output = scribo::debug::char_space_image(input, line); - io::ppm::save(output, argv[2]); + mln::io::ppm::save(output, argv[2]); trace::exiting("main"); } diff --git a/scribo/src/debug/show_info_x_height.cc b/scribo/src/debug/show_info_x_height.cc index e5ff8d2..2997230 100644 --- a/scribo/src/debug/show_info_x_height.cc +++ b/scribo/src/debug/show_info_x_height.cc @@ -1,5 +1,5 @@ -// Copyright (C) 2009, 2010 EPITA Research and Development Laboratory -// (LRDE) +// Copyright (C) 2009, 2010, 2011 EPITA Research and Development +// Laboratory (LRDE) // // This file is part of Olena. // @@ -35,8 +35,8 @@ #include <mln/io/ppm/save.hh> #include <mln/value/rgb8.hh> -#include <mln/value/label_16.hh> +#include <scribo/core/def/lbl_type.hh> #include <scribo/text/extract_lines.hh> #include <scribo/debug/line_info_image.hh> #include <scribo/debug/usage.hh> @@ -66,14 +66,14 @@ int main(int argc, char* argv[]) image2d<bool> input; - io::pbm::load(input, argv[1]); + mln::io::pbm::load(input, argv[1]); - typedef image2d<value::label_16> L; + typedef image2d<scribo::def::lbl_type> L; line_set<L> line = scribo::text::extract_lines(input, c8()); image2d<value::rgb8> output = scribo::debug::line_info_image(input, line); - io::ppm::save(output, argv[2]); + mln::io::ppm::save(output, argv[2]); trace::exiting("main"); } diff --git a/scribo/src/debug/show_text_lines.cc b/scribo/src/debug/show_text_lines.cc index a031072..368f375 100644 --- a/scribo/src/debug/show_text_lines.cc +++ b/scribo/src/debug/show_text_lines.cc @@ -77,7 +77,7 @@ int main(int argc, char* argv[]) util::timer t; t.start(); - typedef image2d<value::int_u16> L; + typedef image2d<scribo::def::lbl_type> L; line_set<L> lines = text::extract_lines(input, c8(), input_seps); t.stop(); diff --git a/scribo/src/primitive/extract/separators_nonvisible.cc b/scribo/src/primitive/extract/separators_nonvisible.cc index 82d4787..d50cc73 100644 --- a/scribo/src/primitive/extract/separators_nonvisible.cc +++ b/scribo/src/primitive/extract/separators_nonvisible.cc @@ -27,7 +27,7 @@ #include <mln/io/pbm/all.hh> #include <mln/data/convert.hh> -#include <scribo/primitive/extract/separators_nonvisible2.hh> +#include <scribo/primitive/extract/separators_nonvisible.hh> #include <scribo/debug/usage.hh> const char *args_desc[][2] = -- 1.5.6.5

1 0

last-svn-commit-786-gb4cc9d5 Regen mk files.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

* scribo/headers.mk, * scribo/tests/unit_test/unit-tests.mk: Regen. * scribo/tests/unit_test/cond_tests_qt: Add scribo/io/xml/load.hh. --- scribo/ChangeLog | 9 +++++++++ scribo/headers.mk | 13 ++++++++++++- scribo/tests/unit_test/cond_tests_qt | 1 + scribo/tests/unit_test/unit-tests.mk | 30 +++++++++++++++++++++++++++--- 4 files changed, 49 insertions(+), 4 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index 9914b73..a300b37 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,3 +1,12 @@ +2011-03-02 Guillaume Lazzara <z(a)lrde.epita.fr> + + Regen mk files. + + * scribo/headers.mk, + * scribo/tests/unit_test/unit-tests.mk: Regen. + + * scribo/tests/unit_test/cond_tests_qt: Add scribo/io/xml/load.hh. + 2011-03-01 Guillaume Lazzara <z(a)lrde.epita.fr> * scribo/primitive/extract/non_text.hh: Remove debug. diff --git a/scribo/headers.mk b/scribo/headers.mk index 7603f8a..0b21a3f 100644 --- a/scribo/headers.mk +++ b/scribo/headers.mk @@ -14,17 +14,22 @@ scribo/binarization/sauvola_ms.hh \ scribo/binarization/sauvola_ms_split.hh \ scribo/binarization/sauvola_threshold_image.hh \ scribo/canvas/integral_browsing.hh \ +scribo/convert/from_base64.hh \ scribo/convert/from_qimage.hh \ +scribo/convert/to_base64.hh \ scribo/core/all.hh \ scribo/core/central_sites.hh \ scribo/core/component_info.hh \ scribo/core/component_set.hh \ scribo/core/concept/dmax_functor.hh \ scribo/core/concept/link_functor.hh \ +scribo/core/concept/serializable.hh \ +scribo/core/concept/serialize_visitor.hh \ scribo/core/def/lbl_type.hh \ scribo/core/document.hh \ scribo/core/erase_objects.hh \ scribo/core/init_integral_image.hh \ +scribo/core/internal/doc_xml_serializer.hh \ scribo/core/line_info.hh \ scribo/core/line_links.hh \ scribo/core/line_set.hh \ @@ -93,6 +98,12 @@ scribo/fun/v2b/label_to_bool.hh \ scribo/fun/v2b/objects_large_filter.hh \ scribo/fun/v2b/objects_small_filter.hh \ scribo/io/text_boxes/save.hh \ +scribo/io/xml/internal/extended_page_xml_visitor.hh \ +scribo/io/xml/internal/full_xml_visitor.hh \ +scribo/io/xml/internal/page_xml_visitor.hh \ +scribo/io/xml/internal/print_box_coords.hh \ +scribo/io/xml/internal/print_page_preambule.hh \ +scribo/io/xml/load.hh \ scribo/io/xml/save.hh \ scribo/make/all.hh \ scribo/make/debug_filename.hh \ @@ -115,7 +126,6 @@ scribo/primitive/extract/all.hh \ scribo/primitive/extract/canvas.hh \ scribo/primitive/extract/cells.hh \ scribo/primitive/extract/components.hh \ -scribo/primitive/extract/elements.hh \ scribo/primitive/extract/horizontal_separators.hh \ scribo/primitive/extract/lines_discontinued.hh \ scribo/primitive/extract/lines_h_discontinued.hh \ @@ -130,6 +140,7 @@ scribo/primitive/extract/lines_v_pattern.hh \ scribo/primitive/extract/lines_v_single.hh \ scribo/primitive/extract/lines_v_thick.hh \ scribo/primitive/extract/lines_v_thick_and_single.hh \ +scribo/primitive/extract/non_text.hh \ scribo/primitive/extract/separators.hh \ scribo/primitive/extract/separators_nonvisible.hh \ scribo/primitive/extract/vertical_separators.hh \ diff --git a/scribo/tests/unit_test/cond_tests_qt b/scribo/tests/unit_test/cond_tests_qt index 5428215..577f9e6 100644 --- a/scribo/tests/unit_test/cond_tests_qt +++ b/scribo/tests/unit_test/cond_tests_qt @@ -1 +1,2 @@ scribo/convert/from_qimage.hh +scribo/io/xml/load.hh diff --git a/scribo/tests/unit_test/unit-tests.mk b/scribo/tests/unit_test/unit-tests.mk index 6a8f29d..71b0b5b 100644 --- a/scribo/tests/unit_test/unit-tests.mk +++ b/scribo/tests/unit_test/unit-tests.mk @@ -19,11 +19,15 @@ endif HAVE_MAGICKXX # Starting a conditional unit test list. if HAVE_QT check_PROGRAMS += \ -scribo_convert_from_qimage +scribo_convert_from_qimage \ +scribo_io_xml_load scribo_convert_from_qimage_CPPFLAGS= ${QT_CPPFLAGS} ${AM_CPPFLAGS} scribo_convert_from_qimage_LDFLAGS= ${QT_LDFLAGS} ${AM_LDFLAGS} scribo_convert_from_qimage_SOURCES = scribo_convert_from_qimage.cc +scribo_io_xml_load_CPPFLAGS= ${QT_CPPFLAGS} ${AM_CPPFLAGS} +scribo_io_xml_load_LDFLAGS= ${QT_LDFLAGS} ${AM_LDFLAGS} +scribo_io_xml_load_SOURCES = scribo_io_xml_load.cc endif HAVE_QT # Starting a conditional unit test list. @@ -79,16 +83,21 @@ scribo_binarization_sauvola_ms \ scribo_binarization_sauvola_ms_split \ scribo_binarization_sauvola_threshold_image \ scribo_canvas_integral_browsing \ +scribo_convert_from_base64 \ +scribo_convert_to_base64 \ scribo_core_all \ scribo_core_central_sites \ scribo_core_component_info \ scribo_core_component_set \ scribo_core_concept_dmax_functor \ scribo_core_concept_link_functor \ +scribo_core_concept_serializable \ +scribo_core_concept_serialize_visitor \ scribo_core_def_lbl_type \ scribo_core_document \ scribo_core_erase_objects \ scribo_core_init_integral_image \ +scribo_core_internal_doc_xml_serializer \ scribo_core_line_info \ scribo_core_line_links \ scribo_core_line_set \ @@ -157,6 +166,11 @@ scribo_fun_v2b_label_to_bool \ scribo_fun_v2b_objects_large_filter \ scribo_fun_v2b_objects_small_filter \ scribo_io_text_boxes_save \ +scribo_io_xml_internal_extended_page_xml_visitor \ +scribo_io_xml_internal_full_xml_visitor \ +scribo_io_xml_internal_page_xml_visitor \ +scribo_io_xml_internal_print_box_coords \ +scribo_io_xml_internal_print_page_preambule \ scribo_io_xml_save \ scribo_make_all \ scribo_make_debug_filename \ @@ -179,7 +193,6 @@ scribo_primitive_extract_all \ scribo_primitive_extract_canvas \ scribo_primitive_extract_cells \ scribo_primitive_extract_components \ -scribo_primitive_extract_elements \ scribo_primitive_extract_horizontal_separators \ scribo_primitive_extract_lines_discontinued \ scribo_primitive_extract_lines_h_discontinued \ @@ -194,6 +207,7 @@ scribo_primitive_extract_lines_v_pattern \ scribo_primitive_extract_lines_v_single \ scribo_primitive_extract_lines_v_thick \ scribo_primitive_extract_lines_v_thick_and_single \ +scribo_primitive_extract_non_text \ scribo_primitive_extract_separators \ scribo_primitive_extract_separators_nonvisible \ scribo_primitive_extract_vertical_separators \ @@ -286,16 +300,21 @@ scribo_binarization_sauvola_ms_SOURCES = scribo_binarization_sauvola_ms.cc scribo_binarization_sauvola_ms_split_SOURCES = scribo_binarization_sauvola_ms_split.cc scribo_binarization_sauvola_threshold_image_SOURCES = scribo_binarization_sauvola_threshold_image.cc scribo_canvas_integral_browsing_SOURCES = scribo_canvas_integral_browsing.cc +scribo_convert_from_base64_SOURCES = scribo_convert_from_base64.cc +scribo_convert_to_base64_SOURCES = scribo_convert_to_base64.cc scribo_core_all_SOURCES = scribo_core_all.cc scribo_core_central_sites_SOURCES = scribo_core_central_sites.cc scribo_core_component_info_SOURCES = scribo_core_component_info.cc scribo_core_component_set_SOURCES = scribo_core_component_set.cc scribo_core_concept_dmax_functor_SOURCES = scribo_core_concept_dmax_functor.cc scribo_core_concept_link_functor_SOURCES = scribo_core_concept_link_functor.cc +scribo_core_concept_serializable_SOURCES = scribo_core_concept_serializable.cc +scribo_core_concept_serialize_visitor_SOURCES = scribo_core_concept_serialize_visitor.cc scribo_core_def_lbl_type_SOURCES = scribo_core_def_lbl_type.cc scribo_core_document_SOURCES = scribo_core_document.cc scribo_core_erase_objects_SOURCES = scribo_core_erase_objects.cc scribo_core_init_integral_image_SOURCES = scribo_core_init_integral_image.cc +scribo_core_internal_doc_xml_serializer_SOURCES = scribo_core_internal_doc_xml_serializer.cc scribo_core_line_info_SOURCES = scribo_core_line_info.cc scribo_core_line_links_SOURCES = scribo_core_line_links.cc scribo_core_line_set_SOURCES = scribo_core_line_set.cc @@ -364,6 +383,11 @@ scribo_fun_v2b_label_to_bool_SOURCES = scribo_fun_v2b_label_to_bool.cc scribo_fun_v2b_objects_large_filter_SOURCES = scribo_fun_v2b_objects_large_filter.cc scribo_fun_v2b_objects_small_filter_SOURCES = scribo_fun_v2b_objects_small_filter.cc scribo_io_text_boxes_save_SOURCES = scribo_io_text_boxes_save.cc +scribo_io_xml_internal_extended_page_xml_visitor_SOURCES = scribo_io_xml_internal_extended_page_xml_visitor.cc +scribo_io_xml_internal_full_xml_visitor_SOURCES = scribo_io_xml_internal_full_xml_visitor.cc +scribo_io_xml_internal_page_xml_visitor_SOURCES = scribo_io_xml_internal_page_xml_visitor.cc +scribo_io_xml_internal_print_box_coords_SOURCES = scribo_io_xml_internal_print_box_coords.cc +scribo_io_xml_internal_print_page_preambule_SOURCES = scribo_io_xml_internal_print_page_preambule.cc scribo_io_xml_save_SOURCES = scribo_io_xml_save.cc scribo_make_all_SOURCES = scribo_make_all.cc scribo_make_debug_filename_SOURCES = scribo_make_debug_filename.cc @@ -386,7 +410,6 @@ scribo_primitive_extract_all_SOURCES = scribo_primitive_extract_all.cc scribo_primitive_extract_canvas_SOURCES = scribo_primitive_extract_canvas.cc scribo_primitive_extract_cells_SOURCES = scribo_primitive_extract_cells.cc scribo_primitive_extract_components_SOURCES = scribo_primitive_extract_components.cc -scribo_primitive_extract_elements_SOURCES = scribo_primitive_extract_elements.cc scribo_primitive_extract_horizontal_separators_SOURCES = scribo_primitive_extract_horizontal_separators.cc scribo_primitive_extract_lines_discontinued_SOURCES = scribo_primitive_extract_lines_discontinued.cc scribo_primitive_extract_lines_h_discontinued_SOURCES = scribo_primitive_extract_lines_h_discontinued.cc @@ -401,6 +424,7 @@ scribo_primitive_extract_lines_v_pattern_SOURCES = scribo_primitive_extract_line scribo_primitive_extract_lines_v_single_SOURCES = scribo_primitive_extract_lines_v_single.cc scribo_primitive_extract_lines_v_thick_SOURCES = scribo_primitive_extract_lines_v_thick.cc scribo_primitive_extract_lines_v_thick_and_single_SOURCES = scribo_primitive_extract_lines_v_thick_and_single.cc +scribo_primitive_extract_non_text_SOURCES = scribo_primitive_extract_non_text.cc scribo_primitive_extract_separators_SOURCES = scribo_primitive_extract_separators.cc scribo_primitive_extract_separators_nonvisible_SOURCES = scribo_primitive_extract_separators_nonvisible.cc scribo_primitive_extract_vertical_separators_SOURCES = scribo_primitive_extract_vertical_separators.cc -- 1.5.6.5

1 0

last-svn-commit-785-g4d0483d scribo/primitive/extract/non_text.hh: Remove debug.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

--- scribo/ChangeLog | 4 ++++ scribo/scribo/primitive/extract/non_text.hh | 21 --------------------- 2 files changed, 4 insertions(+), 21 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index 1eb71a3..9914b73 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,5 +1,9 @@ 2011-03-01 Guillaume Lazzara <z(a)lrde.epita.fr> + * scribo/primitive/extract/non_text.hh: Remove debug. + +2011-03-01 Guillaume Lazzara <z(a)lrde.epita.fr> + Add new debug tools. * scribo/src/debug/Makefile.am: New targets. diff --git a/scribo/scribo/primitive/extract/non_text.hh b/scribo/scribo/primitive/extract/non_text.hh index a017f1f..8528782 100644 --- a/scribo/scribo/primitive/extract/non_text.hh +++ b/scribo/scribo/primitive/extract/non_text.hh @@ -58,10 +58,6 @@ # include <mln/clustering/kmean_rgb.hh> # include <mln/fun/v2v/rgb8_to_rgbn.hh> -# include <mln/util/timer.hh> - -# include <mln/io/pbm/save.hh> - namespace scribo { @@ -139,16 +135,9 @@ namespace scribo image2d<t_rgb5> img_rgb5 = mln::data::transform(doc.image(), t_rgb8_to_rgb5()); - // DEBUG - io::pbm::save(content, "text_area.pbm"); - - mln::util::timer t; - t.start(); img_lbl8 = mln::clustering::kmean_rgb<double,5>((img_rgb5 | pw::value(content)), 3, 10, 10).unmorph_(); data::fill((img_lbl8 | !pw::value(content)).rw(), 0u); - t.stop(); - std::cout << t << std::endl; mln::util::array<unsigned> card = mln::labeling::compute(accu::math::count<value::label_8>(), @@ -178,19 +167,9 @@ namespace scribo data::fill(elts, false); data::fill((elts | (pw::value(img_lbl8) != pw::cst(0))).rw(), true); - // DEBUG - data::fill((elts | (pw::value(doc.line_seps()) != pw::cst(0))).rw(), - false); - // END OF DEBUG - - scribo::def::lbl_type nlabels; elts = filter::components_small(elts, c8(), nlabels, 40); - // DEBUG - io::pbm::save(elts, "elements.pbm"); - // END OF DEBUG - output = primitive::extract::components(elts, c8(), nlabels); } -- 1.5.6.5

1 0

last-svn-commit-784-ga5f7281 Add new debug tools.
by Guillaume Lazzara 29 Mar '11

29 Mar '11

* scribo/src/debug/Makefile.am: New targets. * scribo/src/debug/show_separators.cc, * scribo/src/debug/show_stoppers.cc, * scribo/src/debug/show_text_lines.cc: New. --- scribo/ChangeLog | 10 ++ scribo/src/debug/Makefile.am | 12 ++- .../lines_pattern.cc => debug/show_separators.cc} | 74 +++++++++----- scribo/src/debug/show_stoppers.cc | 104 ++++++++++++++++++++ .../show_text_lines.cc} | 100 +++++++++++-------- 5 files changed, 225 insertions(+), 75 deletions(-) copy scribo/src/{primitive/extract/lines_pattern.cc => debug/show_separators.cc} (50%) create mode 100644 scribo/src/debug/show_stoppers.cc copy scribo/src/{text/pbm_lines_recognition.cc => debug/show_text_lines.cc} (52%) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index c41b0bb..1eb71a3 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,5 +1,15 @@ 2011-03-01 Guillaume Lazzara <z(a)lrde.epita.fr> + Add new debug tools. + + * scribo/src/debug/Makefile.am: New targets. + + * scribo/src/debug/show_separators.cc, + * scribo/src/debug/show_stoppers.cc, + * scribo/src/debug/show_text_lines.cc: New. + +2011-03-01 Guillaume Lazzara <z(a)lrde.epita.fr> + * scribo/text/extract_lines.hh: Update code. 2011-03-01 Guillaume Lazzara <z(a)lrde.epita.fr> diff --git a/scribo/src/debug/Makefile.am b/scribo/src/debug/Makefile.am index d203f5c..cdb1f30 100644 --- a/scribo/src/debug/Makefile.am +++ b/scribo/src/debug/Makefile.am @@ -1,5 +1,5 @@ -# Copyright (C) 2009, 2010 EPITA Research and Development Laboratory -# (LRDE). +# Copyright (C) 2009, 2010, 2011 EPITA Research and Development +# Laboratory (LRDE). # # This file is part of Olena. # @@ -20,6 +20,7 @@ include $(top_srcdir)/scribo/scribo.mk noinst_PROGRAMS = \ show_info_x_height \ show_info_median_inter_characters \ + show_separators \ show_links_left_right_links_validation \ show_links_bbox_h_ratio \ show_links_bbox_overlap \ @@ -43,11 +44,14 @@ noinst_PROGRAMS = \ show_objects_large_small \ show_objects_small \ show_objects_thick \ - show_objects_thin + show_objects_thin \ + show_stoppers \ + show_text_lines show_info_x_height_SOURCES = show_info_x_height.cc show_info_median_inter_characters_SOURCES = show_info_median_inter_characters.cc +show_separators_SOURCES = show_separators.cc show_links_left_right_links_validation_SOURCES = show_links_left_right_links_validation.cc show_links_bbox_h_ratio_SOURCES = show_links_bbox_h_ratio.cc show_links_bbox_overlap_SOURCES = show_links_bbox_overlap.cc @@ -72,6 +76,8 @@ show_objects_large_small_SOURCES = show_objects_large_small.cc show_objects_small_SOURCES = show_objects_small.cc show_objects_thick_SOURCES = show_objects_thick.cc show_objects_thin_SOURCES = show_objects_thin.cc +show_stoppers_SOURCES = show_stoppers.cc +show_text_lines_SOURCES = show_text_lines.cc if HAVE_MAGICKXX diff --git a/scribo/src/primitive/extract/lines_pattern.cc b/scribo/src/debug/show_separators.cc similarity index 50% copy from scribo/src/primitive/extract/lines_pattern.cc copy to scribo/src/debug/show_separators.cc index cb56b36..ef80da6 100644 --- a/scribo/src/primitive/extract/lines_pattern.cc +++ b/scribo/src/debug/show_separators.cc @@ -1,5 +1,4 @@ -// Copyright (C) 2009, 2010 EPITA Research and Development Laboratory -// (LRDE) +// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) // // This file is part of Olena. // @@ -14,33 +13,25 @@ // // You should have received a copy of the GNU General Public License // along with Olena. If not, see <http://www.gnu.org/licenses/>. -// -// As a special exception, you may use this file as part of a free -// software project 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/alias/box2d.hh> +#include <mln/core/alias/neighb2d.hh> #include <mln/io/pbm/all.hh> -#include <mln/pw/all.hh> -#include <mln/data/fill.hh> -#include <mln/core/image/dmorph/image_if.hh> +#include <mln/arith/plus.hh> +#include <mln/labeling/foreground.hh> #include <scribo/primitive/extract/lines_h_pattern.hh> #include <scribo/primitive/extract/lines_v_pattern.hh> #include <scribo/debug/usage.hh> +#include <scribo/debug/save_bboxes_image.hh> const char *args_desc[][2] = { - { "input.pbm", "A binary image." }, - { "length", " Minimum line length. (Common value : 51)" }, + { "length", " Minimum line length. (default: 101)" }, { "delta", " Distance between the object pixel and the background pixel" - "(Common value : 4)" }, + "(default: 4)" }, {0, 0} }; @@ -49,10 +40,10 @@ int main(int argc, char *argv[]) { using namespace mln; - if (argc != 5) + if (argc != 7 && argc != 5) return scribo::debug::usage(argv, - "Extract horizontal lines patterns", - "input.pbm length delta output.pbm", + "Extract horizontal and vertical lines patterns", + "input.pbm output.pbm input_wo_seps.pbm output.ppm length delta", args_desc); trace::entering("main"); @@ -60,19 +51,46 @@ int main(int argc, char *argv[]) image2d<bool> input; io::pbm::load(input, argv[1]); + // Set default option value + unsigned + length = 101, + delta = 4; + if (argc > 5) + { + length = atoi(argv[5]); + delta = atoi(argv[6]); + } + + util::timer t; + t.start(); + image2d<bool> - h_lines = scribo::primitive::extract::lines_h_pattern(input, - atoi(argv[2]), - atoi(argv[3])); + h_lines = scribo::primitive::extract::lines_h_pattern(input, length, delta); image2d<bool> - v_lines = scribo::primitive::extract::lines_v_pattern(input, - atoi(argv[2]), - atoi(argv[3])); + v_lines = scribo::primitive::extract::lines_v_pattern(input, length, delta); + + v_lines += h_lines; + + t.stop(); + std::cout << t << "s" << std::endl; + + // Save binary image. + io::pbm::save(v_lines, argv[2]); - data::fill((v_lines | pw::value(h_lines)).rw(), true); + // Save input without separators + { + image2d<bool> in_wo_seps = duplicate(input); + data::fill((in_wo_seps | pw::value(v_lines)).rw(), false); + io::pbm::save(in_wo_seps, argv[3]); + } - io::pbm::save(v_lines, argv[4]); + // Save bbox image + value::int_u8 nlabels; + image2d<value::int_u8> lbl = labeling::foreground(v_lines, c4(), nlabels); + mln::util::array<box2d> + bbox = labeling::compute(accu::shape::bbox<point2d>(), lbl, nlabels); + scribo::debug::save_bboxes_image(input, bbox, argv[4], literal::red); trace::exiting("main"); } diff --git a/scribo/src/debug/show_stoppers.cc b/scribo/src/debug/show_stoppers.cc new file mode 100644 index 0000000..a108da1 --- /dev/null +++ b/scribo/src/debug/show_stoppers.cc @@ -0,0 +1,104 @@ +// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) +// +// This file is part of Olena. +// +// Olena is free software: you can redistribute it and/or modify it under +// the terms of the GNU General Public License as published by the Free +// Software Foundation, version 2 of the License. +// +// Olena 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 Olena. If not, see <http://www.gnu.org/licenses/>. + +#include <mln/core/image/image2d.hh> +#include <mln/core/alias/box2d.hh> +#include <mln/core/alias/neighb2d.hh> +#include <mln/io/pbm/all.hh> +#include <mln/arith/plus.hh> +#include <mln/labeling/foreground.hh> + +#include <scribo/primitive/extract/lines_h_pattern.hh> +#include <scribo/primitive/extract/lines_v_pattern.hh> +#include <scribo/primitive/extract/separators_nonvisible.hh> +#include <scribo/debug/usage.hh> +#include <scribo/debug/save_bboxes_image.hh> + + +const char *args_desc[][2] = +{ + { "length", " Minimum line length. (default: 101)" }, + { "delta", " Distance between the object pixel and the background pixel" + "(default: 4)" }, + {0, 0} +}; + + +int main(int argc, char *argv[]) +{ + using namespace mln; + + if (argc != 7 && argc != 5) + return scribo::debug::usage(argv, + "Extract horizontal, vertical lines and stoppers", + "input.pbm output.pbm input_wo_seps.pbm output.ppm length delta", + args_desc); + + trace::entering("main"); + + image2d<bool> input; + io::pbm::load(input, argv[1]); + + // Set default option value + unsigned + length = 101, + delta = 4; + if (argc > 5) + { + length = atoi(argv[5]); + delta = atoi(argv[6]); + } + + util::timer t; + + t.start(); + image2d<bool> + h_lines = scribo::primitive::extract::lines_h_pattern(input, length, delta); + + image2d<bool> + v_lines = scribo::primitive::extract::lines_v_pattern(input, length, delta); + + v_lines += h_lines; + + t.stop(); + + // Save input without separators + { + image2d<bool> in_wo_seps = duplicate(input); + data::fill((in_wo_seps | pw::value(v_lines)).rw(), false); + io::pbm::save(in_wo_seps, argv[3]); + } + + // Save bbox image + value::int_u8 nlabels; + image2d<value::int_u8> lbl = labeling::foreground(v_lines, c4(), nlabels); + mln::util::array<box2d> + bbox = labeling::compute(accu::shape::bbox<point2d>(), lbl, nlabels); + scribo::debug::save_bboxes_image(input, bbox, argv[4], literal::red); + + t.resume(); + + // Non visible separators + v_lines += scribo::primitive::extract::separators_nonvisible(input); + + t.stop(); + std::cout << t << "s" << std::endl; + + // Save binary image. + io::pbm::save(v_lines, argv[2]); + + trace::exiting("main"); +} diff --git a/scribo/src/text/pbm_lines_recognition.cc b/scribo/src/debug/show_text_lines.cc similarity index 52% copy from scribo/src/text/pbm_lines_recognition.cc copy to scribo/src/debug/show_text_lines.cc index 3451449..a031072 100644 --- a/scribo/src/text/pbm_lines_recognition.cc +++ b/scribo/src/debug/show_text_lines.cc @@ -1,4 +1,4 @@ -// Copyright (C) 2010 EPITA Research and Development Laboratory (LRDE) +// Copyright (C) 2011 EPITA Research and Development Laboratory (LRDE) // // This file is part of Olena. // @@ -23,15 +23,15 @@ // exception does not however invalidate any other reasons why the // executable file might be covered by the GNU General Public License. -#ifdef HAVE_CONFIG_H -# include <config.h> -#endif - #include <iostream> #include <mln/core/image/image2d.hh> #include <mln/io/pbm/load.hh> #include <mln/value/int_u16.hh> +#include <mln/draw/box_plain.hh> +#include <mln/pw/all.hh> +#include <mln/core/image/dmorph/image_if.hh> +#include <mln/data/convert.hh> #include <scribo/text/recognition.hh> @@ -41,19 +41,15 @@ #include <scribo/core/object_links.hh> #include <scribo/core/object_groups.hh> -#include <scribo/text/merging.hh> -#include <scribo/primitive/extract/components.hh> -#include <scribo/primitive/link/with_single_left_link_dmax_ratio.hh> -#include <scribo/primitive/group/from_single_link.hh> +#include <scribo/text/extract_lines.hh> #include <scribo/io/text_boxes/save.hh> +#include <scribo/debug/save_bboxes_image.hh> + const char *args_desc[][2] = { - { "input.pbm", "A binary image. 'True' for objects, 'False'\ -for the background." }, - { "lang", "Must be set to \"eng\", \"fra\", \"deu\", \"spa\", \"ita\" (Default \"fra\")" }, {0, 0} }; @@ -64,10 +60,10 @@ int main(int argc, char* argv[]) using namespace scribo; using namespace mln; - if (argc != 2 && argc != 3 && argc != 4) + if (argc != 6) return scribo::debug::usage(argv, - "Text recognition", - "input.pbm [lang] [file]", + "Show text lines", + "input.pbm input_seps.pbm out.ppm out.pbm comps.pbm", args_desc); trace::entering("main"); @@ -75,37 +71,53 @@ int main(int argc, char* argv[]) image2d<bool> input; mln::io::pbm::load(input, argv[1]); - typedef image2d<value::int_u16> L; - value::int_u16 ncomps; - component_set<L> - comps = scribo::primitive::extract::components(input, c8(), ncomps); - std::cout << "ncomps = " << ncomps << std::endl; - - object_links<L> - links = scribo::primitive::link::with_single_left_link_dmax_ratio(comps, 2); - - object_groups<L> - groups = scribo::primitive::group::from_single_link(links); + image2d<bool> input_seps; + mln::io::pbm::load(input_seps, argv[2]); - line_set<L> lines(groups); - lines = text::merging(lines); + util::timer t; + t.start(); - std::string str = argv[2]; - const char *lang; - if (argc < 3 || (str != "eng" && str != "fra" - && str != "deu" && str != "spa" - && str != "ita")) - lang = "fra"; - else - lang = argv[2]; - - char *output = 0; - if (argc >= 4) - output = argv[3]; - - scribo::text::recognition(lines, lang); - - scribo::io::text_boxes::save(lines, output); + typedef image2d<value::int_u16> L; + line_set<L> lines = text::extract_lines(input, c8(), input_seps); + + t.stop(); + std::cout << t << "s" << std::endl; + + // Color image with line bboxes. + { + image2d<value::rgb8> output = data::convert(value::rgb8(), input); + for_all_lines(l, lines) + if (lines(l).is_textline()) + mln::draw::box(output, lines(l).bbox(), literal::red); + mln::io::ppm::save(output, argv[3]); + } + + // Binary image with line bboxes. + { + image2d<bool> output; + initialize(output, input); + data::fill(output, false); + for_all_lines(l, lines) + if (lines(l).is_textline()) + mln::draw::box_plain(output, lines(l).bbox(), true); + mln::io::pbm::save(output, argv[4]); + } + + { + image2d<bool> output; + initialize(output, input); + data::fill(output, false); + for_all_lines(l, lines) + if (lines(l).is_textline()) + { + const util::array<component_id_t>& comps = lines(l).components(); + const L& lbl = lines.components().labeled_image(); + for_all_elements(c, comps) + data::fill((output | lines.components()(comps(c)).bbox()).rw(), + ((input | lines.components()(comps(c)).bbox()) | (pw::value(lbl) == comps(c)))); + } + mln::io::pbm::save(output, argv[5]); + } trace::exiting("main"); } -- 1.5.6.5

1 0