/**********************************************************************
* run_cube_combiner
*
* Iterates through tesseract's results and calls cube on each word,
* combining the results with the existing tesseract result.
**********************************************************************/
void Tesseract::run_cube_combiner(PAGE_RES *page_res) {
if (page_res == NULL || tess_cube_combiner_ == NULL)
return;
PAGE_RES_IT page_res_it(page_res);
// Iterate through the word results and call cube on each word.
for (page_res_it.restart_page(); page_res_it.word () != NULL;
page_res_it.forward()) {
BLOCK* block = page_res_it.block()->block;
if (block->poly_block() != NULL && !block->poly_block()->IsText())
continue; // Don't deal with non-text blocks.
WERD_RES* word = page_res_it.word();
// Skip cube entirely if tesseract's certainty is greater than threshold.
int combiner_run_thresh = convert_prob_to_tess_certainty(
cube_cntxt_->Params()->CombinerRunThresh());
if (word->best_choice->certainty() >= combiner_run_thresh) {
continue;
}
// Use the same language as Tesseract used for the word.
Tesseract* lang_tess = word->tesseract;
// Setup a trial WERD_RES in which to classify with cube.
WERD_RES cube_word;
cube_word.InitForRetryRecognition(*word);
cube_word.SetupForRecognition(lang_tess->unicharset, this, BestPix(),
OEM_CUBE_ONLY,
NULL, false, false, false,
page_res_it.row()->row,
page_res_it.block()->block);
CubeObject *cube_obj = lang_tess->cube_recognize_word(
page_res_it.block()->block, &cube_word);
if (cube_obj != NULL)
lang_tess->cube_combine_word(cube_obj, &cube_word, word);
delete cube_obj;
}
}
/**********************************************************************
* run_cube_combiner
*
* Iterates through tesseract's results and calls cube on each word,
* combining the results with the existing tesseract result.
**********************************************************************/
void Tesseract::run_cube_combiner(PAGE_RES *page_res) {
if (page_res == NULL || tess_cube_combiner_ == NULL)
return;
PAGE_RES_IT page_res_it(page_res);
// Iterate through the word results and call cube on each word.
for (page_res_it.restart_page(); page_res_it.word () != NULL;
page_res_it.forward()) {
WERD_RES* word = page_res_it.word();
// Skip cube entirely if tesseract's certainty is greater than threshold.
int combiner_run_thresh = convert_prob_to_tess_certainty(
cube_cntxt_->Params()->CombinerRunThresh());
if (word->best_choice->certainty() >= combiner_run_thresh) {
continue;
}
// Use the same language as Tesseract used for the word.
Tesseract* lang_tess = word->tesseract;
// Setup a trial WERD_RES in which to classify with cube.
WERD_RES cube_word;
cube_word.InitForRetryRecognition(*word);
CubeObject *cube_obj = lang_tess->cube_recognize_word(
page_res_it.block()->block, &cube_word);
if (cube_obj != NULL)
lang_tess->cube_combine_word(cube_obj, &cube_word, word);
delete cube_obj;
}
}
/**********************************************************************
* cube_recognize
*
* Call cube on the current word, and write the result to word.
* Sets up a fake result and returns false if something goes wrong.
**********************************************************************/
bool Tesseract::cube_recognize(CubeObject *cube_obj, BLOCK* block,
WERD_RES *word) {
// Run cube
WordAltList *cube_alt_list = cube_obj->RecognizeWord();
if (!cube_alt_list || cube_alt_list->AltCount() <= 0) {
if (cube_debug_level > 0) {
tprintf("Cube returned nothing for word at:");
word->word->bounding_box().print();
}
word->SetupFake(unicharset);
return false;
}
// Get cube's best result and its probability, mapped to tesseract's
// certainty range
char_32 *cube_best_32 = cube_alt_list->Alt(0);
double cube_prob = CubeUtils::Cost2Prob(cube_alt_list->AltCost(0));
float cube_certainty = convert_prob_to_tess_certainty(cube_prob);
string cube_best_str;
CubeUtils::UTF32ToUTF8(cube_best_32, &cube_best_str);
// Retrieve Cube's character bounding boxes and CharSamples,
// corresponding to the most recent call to RecognizeWord().
Boxa *char_boxes = NULL;
CharSamp **char_samples = NULL;;
int num_chars;
if (!extract_cube_state(cube_obj, &num_chars, &char_boxes, &char_samples)
&& cube_debug_level > 0) {
tprintf("Cube WARNING (Tesseract::cube_recognize): Cannot extract "
"cube state.\n");
word->SetupFake(unicharset);
return false;
}
// Convert cube's character bounding boxes to a BoxWord.
BoxWord cube_box_word;
TBOX tess_word_box = word->word->bounding_box();
if (word->denorm.block() != NULL)
tess_word_box.rotate(word->denorm.block()->re_rotation());
bool box_word_success = create_cube_box_word(char_boxes, num_chars,
tess_word_box,
&cube_box_word);
boxaDestroy(&char_boxes);
if (!box_word_success) {
if (cube_debug_level > 0) {
tprintf("Cube WARNING (Tesseract::cube_recognize): Could not "
"create cube BoxWord\n");
}
word->SetupFake(unicharset);
return false;
}
// Fill tesseract result's fields with cube results
fill_werd_res(cube_box_word, cube_best_str.c_str(), word);
// Create cube's best choice.
BLOB_CHOICE** choices = new BLOB_CHOICE*[num_chars];
for (int i = 0; i < num_chars; ++i) {
UNICHAR_ID uch_id =
cube_cntxt_->CharacterSet()->UnicharID(char_samples[i]->StrLabel());
choices[i] = new BLOB_CHOICE(uch_id, -cube_certainty, cube_certainty,
-1, -1, 0, 0, 0, 0, BCC_STATIC_CLASSIFIER);
}
word->FakeClassifyWord(num_chars, choices);
// within a word, cube recognizes the word in reading order.
word->best_choice->set_unichars_in_script_order(true);
delete [] choices;
delete [] char_samples;
// Some sanity checks
ASSERT_HOST(word->best_choice->length() == word->reject_map.length());
if (cube_debug_level || classify_debug_level) {
tprintf("Cube result: %s r=%g, c=%g\n",
word->best_choice->unichar_string().string(),
word->best_choice->rating(),
word->best_choice->certainty());
}
return true;
}
/**********************************************************************
* cube_recognize
*
* Call cube on the current word, and write the result to word.
* Sets up a fake result and returns false if something goes wrong.
**********************************************************************/
bool Tesseract::cube_recognize(CubeObject *cube_obj, BLOCK* block,
WERD_RES *word) {
if (!word->SetupForCubeRecognition(unicharset, this, block)) {
return false; // Graphics block.
}
// Run cube
WordAltList *cube_alt_list = cube_obj->RecognizeWord();
if (!cube_alt_list || cube_alt_list->AltCount() <= 0) {
if (cube_debug_level > 0) {
tprintf("Cube returned nothing for word at:");
word->word->bounding_box().print();
}
word->SetupFake(unicharset);
return false;
}
// Get cube's best result and its probability, mapped to tesseract's
// certainty range
char_32 *cube_best_32 = cube_alt_list->Alt(0);
double cube_prob = CubeUtils::Cost2Prob(cube_alt_list->AltCost(0));
float cube_certainty = convert_prob_to_tess_certainty(cube_prob);
std::string cube_best_str;
CubeUtils::UTF32ToUTF8(cube_best_32, &cube_best_str);
// Retrieve Cube's character bounding boxes and CharSamples,
// corresponding to the most recent call to RecognizeWord().
Boxa *char_boxes = NULL;
CharSamp **char_samples = NULL;;
int num_chars;
if (!extract_cube_state(cube_obj, &num_chars, &char_boxes, &char_samples)
&& cube_debug_level > 0) {
tprintf("Cube WARNING (Tesseract::cube_recognize): Cannot extract "
"cube state.\n");
word->SetupFake(unicharset);
return false;
}
// Convert cube's character bounding boxes to a BoxWord.
BoxWord cube_box_word;
TBOX tess_word_box = word->word->bounding_box();
if (word->denorm.block() != NULL)
tess_word_box.rotate(word->denorm.block()->re_rotation());
bool box_word_success = create_cube_box_word(char_boxes, num_chars,
tess_word_box,
&cube_box_word);
boxaDestroy(&char_boxes);
if (!box_word_success) {
if (cube_debug_level > 0) {
tprintf("Cube WARNING (Tesseract::cube_recognize): Could not "
"create cube BoxWord\n");
}
word->SetupFake(unicharset);
return false;
}
// Create cube's best choice.
WERD_CHOICE* cube_werd_choice = create_werd_choice(
char_samples, num_chars, cube_best_str.c_str(), cube_certainty,
unicharset, cube_cntxt_->CharacterSet());
delete []char_samples;
if (!cube_werd_choice) {
if (cube_debug_level > 0) {
tprintf("Cube WARNING (Tesseract::cube_recognize): Could not "
"create cube WERD_CHOICE\n");
}
word->SetupFake(unicharset);
return false;
}
if (cube_debug_level || classify_debug_level) {
tprintf("Cube result: %s r=%g, c=%g\n",
cube_werd_choice->unichar_string().string(),
cube_werd_choice->rating(),
cube_werd_choice->certainty());
}
// Fill tesseract result's fields with cube results
fill_werd_res(cube_box_word, cube_werd_choice, cube_best_str.c_str(), word);
return true;
}