/**
* This program reads in a text file consisting of feature
* samples from a training page in the following format:
* @verbatim
FontName UTF8-char-str xmin ymin xmax ymax page-number
NumberOfFeatureTypes(N)
FeatureTypeName1 NumberOfFeatures(M)
Feature1
...
FeatureM
FeatureTypeName2 NumberOfFeatures(M)
Feature1
...
FeatureM
...
FeatureTypeNameN NumberOfFeatures(M)
Feature1
...
FeatureM
FontName CharName ...
@endverbatim
* The result of this program is a binary inttemp file used by
* the OCR engine.
* @param argc number of command line arguments
* @param argv array of command line arguments
* @return none
* @note Exceptions: none
* @note History: Fri Aug 18 08:56:17 1989, DSJ, Created.
* @note History: Mon May 18 1998, Christy Russson, Revistion started.
*/
int main (int argc, char **argv) {
ParseArguments(&argc, &argv);
ShapeTable* shape_table = NULL;
STRING file_prefix;
// Load the training data.
MasterTrainer* trainer = tesseract::LoadTrainingData(argc, argv,
false,
&shape_table,
&file_prefix);
if (trainer == NULL)
return 1; // Failed.
// Setup an index mapping from the shapes in the shape table to the classes
// that will be trained. In keeping with the original design, each shape
// with the same list of unichars becomes a different class and the configs
// represent the different combinations of fonts.
IndexMapBiDi config_map;
SetupConfigMap(shape_table, &config_map);
WriteShapeTable(file_prefix, *shape_table);
// If the shape_table is flat, then either we didn't run shape clustering, or
// it did nothing, so we just output the trainer's unicharset.
// Otherwise shape_set will hold a fake unicharset with an entry for each
// shape in the shape table, and we will output that instead.
UNICHARSET shape_set;
const UNICHARSET* unicharset = &trainer->unicharset();
// If we ran shapeclustering (and it worked) then at least one shape will
// have multiple unichars, so we have to build a fake unicharset.
if (shape_table->AnyMultipleUnichars()) {
unicharset = &shape_set;
// Now build a fake unicharset for the compact shape space to keep the
// output modules happy that we are doing things correctly.
int num_shapes = config_map.CompactSize();
for (int s = 0; s < num_shapes; ++s) {
char shape_label[kMaxShapeLabelLength + 1];
snprintf(shape_label, kMaxShapeLabelLength, "sh%04d", s);
shape_set.unichar_insert(shape_label);
}
}
// Now train each config separately.
int num_configs = shape_table->NumShapes();
LIST mf_classes = NIL_LIST;
for (int s = 0; s < num_configs; ++s) {
int unichar_id, font_id;
if (unicharset == &shape_set) {
// Using fake unichar_ids from the config_map/shape_set.
unichar_id = config_map.SparseToCompact(s);
} else {
// Get the real unichar_id from the shape table/unicharset.
shape_table->GetFirstUnicharAndFont(s, &unichar_id, &font_id);
}
const char* class_label = unicharset->id_to_unichar(unichar_id);
mf_classes = ClusterOneConfig(s, class_label, mf_classes, *shape_table,
trainer);
}
STRING inttemp_file = file_prefix;
inttemp_file += "inttemp";
STRING pffmtable_file = file_prefix;
pffmtable_file += "pffmtable";
CLASS_STRUCT* float_classes = SetUpForFloat2Int(*unicharset, mf_classes);
// Now write the inttemp and pffmtable.
trainer->WriteInttempAndPFFMTable(trainer->unicharset(), *unicharset,
*shape_table, float_classes,
inttemp_file.string(),
pffmtable_file.string());
for (int c = 0; c < unicharset->size(); ++c) {
FreeClassFields(&float_classes[c]);
}
delete [] float_classes;
FreeLabeledClassList(mf_classes);
delete trainer;
delete shape_table;
printf("Done!\n");
if (!FLAGS_test_ch.empty()) {
// If we are displaying debug window(s), wait for the user to look at them.
printf("Hit return to exit...\n");
while (getchar() != '\n');
}
return 0;
} /* main */
/*---------------------------------------------------------------------------*/
int main (int argc, char **argv) {
/*
** Parameters:
** argc number of command line arguments
** argv array of command line arguments
** Globals: none
** Operation:
** This program reads in a text file consisting of feature
** samples from a training page in the following format:
**
** FontName UTF8-char-str xmin ymin xmax ymax page-number
** NumberOfFeatureTypes(N)
** FeatureTypeName1 NumberOfFeatures(M)
** Feature1
** ...
** FeatureM
** FeatureTypeName2 NumberOfFeatures(M)
** Feature1
** ...
** FeatureM
** ...
** FeatureTypeNameN NumberOfFeatures(M)
** Feature1
** ...
** FeatureM
** FontName CharName ...
**
** The result of this program is a binary inttemp file used by
** the OCR engine.
** Return: none
** Exceptions: none
** History: Fri Aug 18 08:56:17 1989, DSJ, Created.
** Mon May 18 1998, Christy Russson, Revistion started.
*/
ParseArguments(&argc, &argv);
ShapeTable* shape_table = NULL;
STRING file_prefix;
// Load the training data.
MasterTrainer* trainer = tesseract::LoadTrainingData(argc, argv,
false,
&shape_table,
&file_prefix);
if (trainer == NULL)
return 1; // Failed.
// Setup an index mapping from the shapes in the shape table to the classes
// that will be trained. In keeping with the original design, each shape
// with the same list of unichars becomes a different class and the configs
// represent the different combinations of fonts.
IndexMapBiDi config_map;
SetupConfigMap(shape_table, &config_map);
WriteShapeTable(file_prefix, *shape_table);
// If the shape_table is flat, then either we didn't run shape clustering, or
// it did nothing, so we just output the trainer's unicharset.
// Otherwise shape_set will hold a fake unicharset with an entry for each
// shape in the shape table, and we will output that instead.
UNICHARSET shape_set;
const UNICHARSET* unicharset = &trainer->unicharset();
// If we ran shapeclustering (and it worked) then at least one shape will
// have multiple unichars, so we have to build a fake unicharset.
if (shape_table->AnyMultipleUnichars()) {
unicharset = &shape_set;
// Now build a fake unicharset for the compact shape space to keep the
// output modules happy that we are doing things correctly.
int num_shapes = config_map.CompactSize();
for (int s = 0; s < num_shapes; ++s) {
char shape_label[kMaxShapeLabelLength + 1];
snprintf(shape_label, kMaxShapeLabelLength, "sh%04d", s);
shape_set.unichar_insert(shape_label);
}
}
// Now train each config separately.
int num_configs = shape_table->NumShapes();
LIST mf_classes = NIL_LIST;
for (int s = 0; s < num_configs; ++s) {
int unichar_id, font_id;
if (unicharset == &shape_set) {
// Using fake unichar_ids from the config_map/shape_set.
unichar_id = config_map.SparseToCompact(s);
} else {
// Get the real unichar_id from the shape table/unicharset.
shape_table->GetFirstUnicharAndFont(s, &unichar_id, &font_id);
}
const char* class_label = unicharset->id_to_unichar(unichar_id);
mf_classes = ClusterOneConfig(s, class_label, mf_classes, *shape_table,
trainer);
}
STRING inttemp_file = file_prefix;
inttemp_file += "inttemp";
STRING pffmtable_file = file_prefix;
pffmtable_file += "pffmtable";
CLASS_STRUCT* float_classes = SetUpForFloat2Int(*unicharset, mf_classes);
// Now write the inttemp and pffmtable.
trainer->WriteInttempAndPFFMTable(trainer->unicharset(), *unicharset,
*shape_table, float_classes,
inttemp_file.string(),
pffmtable_file.string());
delete [] float_classes;
FreeLabeledClassList(mf_classes);
delete trainer;
delete shape_table;
printf("Done!\n");
if (!FLAGS_test_ch.empty()) {
// If we are displaying debug window(s), wait for the user to look at them.
printf("Hit return to exit...\n");
while (getchar() != '\n');
}
return 0;
} /* main */