// Make a text string from the internal data structures.
// The input page_res is deleted.
char* TessBaseAPI::TesseractToText(PAGE_RES* page_res) {
if (page_res != NULL) {
int total_length = TextLength(page_res);
PAGE_RES_IT page_res_it(page_res);
char* result = new char[total_length];
char* ptr = result;
for (page_res_it.restart_page(); page_res_it.word () != NULL;
page_res_it.forward()) {
WERD_RES *word = page_res_it.word();
WERD_CHOICE* choice = word->best_choice;
if (choice != NULL) {
strcpy(ptr, choice->string().string());
ptr += strlen(ptr);
if (word->word->flag(W_EOL))
*ptr++ = '\n';
else
*ptr++ = ' ';
}
}
*ptr++ = '\n';
*ptr = '\0';
delete page_res;
return result;
}
return NULL;
}
/**********************************************************************
* 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;
}
}
// Extract the OCR results, costs (penalty points for uncertainty),
// and the bounding boxes of the characters.
static void extract_result(ELIST_ITERATOR *out,
PAGE_RES* page_res) {
PAGE_RES_IT page_res_it(page_res);
int word_count = 0;
while (page_res_it.word() != NULL) {
WERD_RES *word = page_res_it.word();
const char *str = word->best_choice->string().string();
const char *len = word->best_choice->lengths().string();
if (word_count)
add_space(out);
TBOX bln_rect;
PBLOB_LIST *blobs = word->outword->blob_list();
PBLOB_IT it(blobs);
int n = strlen(len);
TBOX** boxes_to_fix = new TBOX*[n];
for (int i = 0; i < n; i++) {
PBLOB *blob = it.data();
TBOX current = blob->bounding_box();
bln_rect = bln_rect.bounding_union(current);
TESS_CHAR *tc = new TESS_CHAR(rating_to_cost(word->best_choice->rating()),
str, *len);
tc->box = current;
boxes_to_fix[i] = &tc->box;
out->add_after_then_move(tc);
it.forward();
str += *len;
len++;
}
// Find the word bbox before normalization.
// Here we can't use the C_BLOB bboxes directly,
// since connected letters are not yet cut.
TBOX real_rect = word->word->bounding_box();
// Denormalize boxes by transforming the bbox of the whole bln word
// into the denorm bbox (`real_rect') of the whole word.
double x_stretch = double(real_rect.width()) / bln_rect.width();
double y_stretch = double(real_rect.height()) / bln_rect.height();
for (int j = 0; j < n; j++) {
TBOX *box = boxes_to_fix[j];
int x0 = int(real_rect.left() +
x_stretch * (box->left() - bln_rect.left()) + 0.5);
int x1 = int(real_rect.left() +
x_stretch * (box->right() - bln_rect.left()) + 0.5);
int y0 = int(real_rect.bottom() +
y_stretch * (box->bottom() - bln_rect.bottom()) + 0.5);
int y1 = int(real_rect.bottom() +
y_stretch * (box->top() - bln_rect.bottom()) + 0.5);
*box = TBOX(ICOORD(x0, y0), ICOORD(x1, y1));
}
delete [] boxes_to_fix;
page_res_it.forward();
word_count++;
}
}
/**
* @name process_selected_words()
*
* Walk the current block list applying the specified word processor function
* to each word that overlaps the selection_box.
*/
void Tesseract::process_selected_words(
PAGE_RES* page_res, // blocks to check
TBOX & selection_box,
BOOL8(tesseract::Tesseract::*word_processor)(PAGE_RES_IT* pr_it)) {
for (PAGE_RES_IT page_res_it(page_res); page_res_it.word() != NULL;
page_res_it.forward()) {
WERD* word = page_res_it.word()->word;
if (word->bounding_box().overlap(selection_box)) {
if (!(this->*word_processor)(&page_res_it))
return;
}
}
}
// Make a text string from the internal data structures.
// The input page_res is deleted.
// The text string takes the form of a box file as needed for training.
char* TessBaseAPI::TesseractToBoxText(PAGE_RES* page_res,
int left, int bottom) {
if (page_res != NULL) {
int total_length = TextLength(page_res) * kMaxCharsPerChar;
PAGE_RES_IT page_res_it(page_res);
char* result = new char[total_length];
char* ptr = result;
for (page_res_it.restart_page(); page_res_it.word () != NULL;
page_res_it.forward()) {
WERD_RES *word = page_res_it.word();
ptr += ConvertWordToBoxText(word,page_res_it.row(),left, bottom, ptr);
}
*ptr = '\0';
delete page_res;
return result;
}
return NULL;
}
// Return the maximum length that the output text string might occupy.
int TessBaseAPI::TextLength(PAGE_RES* page_res) {
PAGE_RES_IT page_res_it(page_res);
int total_length = 2;
// Iterate over the data structures to extract the recognition result.
for (page_res_it.restart_page(); page_res_it.word () != NULL;
page_res_it.forward()) {
WERD_RES *word = page_res_it.word();
WERD_CHOICE* choice = word->best_choice;
if (choice != NULL) {
total_length += choice->string().length() + 1;
for (int i = 0; i < word->reject_map.length(); ++i) {
if (word->reject_map[i].rejected())
++total_length;
}
}
}
return total_length;
}
char* TessBaseAPI::TesseractToUNLV(PAGE_RES* page_res) {
bool tilde_crunch_written = false;
bool last_char_was_newline = true;
bool last_char_was_tilde = false;
if (page_res != NULL) {
int total_length = TextLength(page_res);
PAGE_RES_IT page_res_it(page_res);
char* result = new char[total_length];
char* ptr = result;
for (page_res_it.restart_page(); page_res_it.word () != NULL;
page_res_it.forward()) {
WERD_RES *word = page_res_it.word();
// Process the current word.
if (word->unlv_crunch_mode != CR_NONE) {
if (word->unlv_crunch_mode != CR_DELETE &&
(!tilde_crunch_written ||
(word->unlv_crunch_mode == CR_KEEP_SPACE &&
word->word->space () > 0 &&
!word->word->flag (W_FUZZY_NON) &&
!word->word->flag (W_FUZZY_SP)))) {
if (!word->word->flag (W_BOL) &&
word->word->space () > 0 &&
!word->word->flag (W_FUZZY_NON) &&
!word->word->flag (W_FUZZY_SP)) {
/* Write a space to separate from preceeding good text */
*ptr++ = ' ';
last_char_was_tilde = false;
}
if (!last_char_was_tilde) {
// Write a reject char.
last_char_was_tilde = true;
*ptr++ = kUnrecognized;
tilde_crunch_written = true;
last_char_was_newline = false;
}
}
} else {
// NORMAL PROCESSING of non tilde crunched words.
tilde_crunch_written = false;
if (last_char_was_tilde &&
word->word->space () == 0 &&
(word->best_choice->string ()[0] == ' ')) {
/* Prevent adjacent tilde across words - we know that adjacent tildes within
words have been removed */
char* p = (char *) word->best_choice->string().string ();
strcpy (p, p + 1); //shuffle up
p = (char *) word->best_choice->lengths().string ();
strcpy (p, p + 1); //shuffle up
word->reject_map.remove_pos (0);
PBLOB_IT blob_it = word->outword->blob_list ();
delete blob_it.extract (); //get rid of reject blob
}
if (word->word->flag(W_REP_CHAR) && tessedit_consistent_reps)
ensure_rep_chars_are_consistent(word);
set_unlv_suspects(word);
const char* wordstr = word->best_choice->string().string();
if (wordstr[0] != 0) {
if (!last_char_was_newline)
*ptr++ = ' ';
else
last_char_was_newline = false;
int offset = 0;
const STRING& lengths = word->best_choice->lengths();
int length = lengths.length();
for (int i = 0; i < length; offset += lengths[i++]) {
if (wordstr[offset] == ' ' ||
wordstr[offset] == '~' ||
wordstr[offset] == '|') {
*ptr++ = kUnrecognized;
last_char_was_tilde = true;
} else {
if (word->reject_map[i].rejected())
*ptr++ = '^';
UNICHAR ch(wordstr + offset, lengths[i]);
int uni_ch = ch.first_uni();
for (int j = 0; kUniChs[j] != 0; ++j) {
if (kUniChs[j] == uni_ch) {
uni_ch = kLatinChs[j];
break;
}
}
if (uni_ch <= 0xff) {
*ptr++ = static_cast<char>(uni_ch);
last_char_was_tilde = false;
} else {
*ptr++ = kUnrecognized;
last_char_was_tilde = true;
}
}
}
}
}
if (word->word->flag(W_EOL) && !last_char_was_newline) {
/* Add a new line output */
*ptr++ = '\n';
tilde_crunch_written = false;
last_char_was_newline = true;
last_char_was_tilde = false;
}
}
*ptr++ = '\n';
*ptr = '\0';
delete page_res;
return result;
}
return NULL;
}
/// Gather consecutive blobs that match the given box into the best_state
/// and corresponding correct_text.
///
/// Fights over which box owns which blobs are settled by pre-chopping and
/// applying the blobs to box or next_box with the least non-overlap.
/// @return false if the box was in error, which can only be caused by
/// failing to find an appropriate blob for a box.
///
/// This means that occasionally, blobs may be incorrectly segmented if the
/// chopper fails to find a suitable chop point.
bool Tesseract::ResegmentCharBox(PAGE_RES* page_res, const TBOX *prev_box,
const TBOX& box, const TBOX& next_box,
const char* correct_text) {
if (applybox_debug > 1) {
tprintf("\nAPPLY_BOX: in ResegmentCharBox() for %s\n", correct_text);
}
PAGE_RES_IT page_res_it(page_res);
WERD_RES* word_res;
for (word_res = page_res_it.word(); word_res != NULL;
word_res = page_res_it.forward()) {
if (!word_res->box_word->bounding_box().major_overlap(box))
continue;
if (applybox_debug > 1) {
tprintf("Checking word box:");
word_res->box_word->bounding_box().print();
}
int word_len = word_res->box_word->length();
for (int i = 0; i < word_len; ++i) {
TBOX char_box = TBOX();
int blob_count = 0;
for (blob_count = 0; i + blob_count < word_len; ++blob_count) {
TBOX blob_box = word_res->box_word->BlobBox(i + blob_count);
if (!blob_box.major_overlap(box))
break;
if (word_res->correct_text[i + blob_count].length() > 0)
break; // Blob is claimed already.
double current_box_miss_metric = BoxMissMetric(blob_box, box);
double next_box_miss_metric = BoxMissMetric(blob_box, next_box);
if (applybox_debug > 2) {
tprintf("Checking blob:");
blob_box.print();
tprintf("Current miss metric = %g, next = %g\n",
current_box_miss_metric, next_box_miss_metric);
}
if (current_box_miss_metric > next_box_miss_metric)
break; // Blob is a better match for next box.
char_box += blob_box;
}
if (blob_count > 0) {
if (applybox_debug > 1) {
tprintf("Index [%d, %d) seem good.\n", i, i + blob_count);
}
if (!char_box.almost_equal(box, 3) &&
(box.x_gap(next_box) < -3 ||
(prev_box != NULL && prev_box->x_gap(box) < -3))) {
return false;
}
// We refine just the box_word, best_state and correct_text here.
// The rebuild_word is made in TidyUp.
// blob_count blobs are put together to match the box. Merge the
// box_word boxes, save the blob_count in the state and the text.
word_res->box_word->MergeBoxes(i, i + blob_count);
word_res->best_state[i] = blob_count;
word_res->correct_text[i] = correct_text;
if (applybox_debug > 2) {
tprintf("%d Blobs match: blob box:", blob_count);
word_res->box_word->BlobBox(i).print();
tprintf("Matches box:");
box.print();
tprintf("With next box:");
next_box.print();
}
// Eliminated best_state and correct_text entries for the consumed
// blobs.
for (int j = 1; j < blob_count; ++j) {
word_res->best_state.remove(i + 1);
word_res->correct_text.remove(i + 1);
}
// Assume that no box spans multiple source words, so we are done with
// this box.
if (applybox_debug > 1) {
tprintf("Best state = ");
for (int j = 0; j < word_res->best_state.size(); ++j) {
tprintf("%d ", word_res->best_state[j]);
}
tprintf("\n");
tprintf("Correct text = [[ ");
for (int j = 0; j < word_res->correct_text.size(); ++j) {
tprintf("%s ", word_res->correct_text[j].string());
}
tprintf("]]\n");
}
return true;
}
}
}
if (applybox_debug > 0) {
tprintf("FAIL!\n");
}
return false; // Failure.
}