/**********************************************************************
* get_piece_rating
*
* Check to see if this piece has already been classified. If it has
* return that rating. Otherwise build the piece from the smaller
* pieces, classify it, store the rating for later, and take the piece
* apart again.
**********************************************************************/
CHOICES get_piece_rating(MATRIX ratings,
TBLOB *blobs,
SEAMS seams,
INT16 start,
INT16 end,
INT32 fx,
STATE *this_state,
STATE *best_state,
INT32 pass,
INT32 blob_index) {
CHOICES choices;
choices = matrix_get (ratings, start, end);
if (choices == NOT_CLASSIFIED) {
choices =
classify_piece(blobs,
seams,
start,
end,
fx,
this_state,
best_state,
pass,
blob_index);
matrix_put(ratings, start, end, choices);
}
return (choices);
}
void Wordrec::ProcessSegSearchPainPoint(
float pain_point_priority,
const MATRIX_COORD &pain_point, const char* pain_point_type,
GenericVector<SegSearchPending>* pending, WERD_RES *word_res,
LMPainPoints *pain_points, BlamerBundle *blamer_bundle) {
if (segsearch_debug_level > 0) {
tprintf("Classifying pain point %s priority=%.4f, col=%d, row=%d\n",
pain_point_type, pain_point_priority,
pain_point.col, pain_point.row);
}
ASSERT_HOST(pain_points != NULL);
MATRIX *ratings = word_res->ratings;
// Classify blob [pain_point.col pain_point.row]
if (!pain_point.Valid(*ratings)) {
ratings->IncreaseBandSize(pain_point.row + 1 - pain_point.col);
}
ASSERT_HOST(pain_point.Valid(*ratings));
BLOB_CHOICE_LIST *classified = classify_piece(word_res->seam_array,
pain_point.col, pain_point.row,
pain_point_type,
word_res->chopped_word,
blamer_bundle);
BLOB_CHOICE_LIST *lst = ratings->get(pain_point.col, pain_point.row);
if (lst == NULL) {
ratings->put(pain_point.col, pain_point.row, classified);
} else {
// We can not delete old BLOB_CHOICEs, since they might contain
// ViterbiStateEntries that are parents of other "active" entries.
// Thus if the matrix cell already contains classifications we add
// the new ones to the beginning of the list.
BLOB_CHOICE_IT it(lst);
it.add_list_before(classified);
delete classified; // safe to delete, since empty after add_list_before()
classified = NULL;
}
if (segsearch_debug_level > 0) {
print_ratings_list("Updated ratings matrix with a new entry:",
ratings->get(pain_point.col, pain_point.row),
getDict().getUnicharset());
ratings->print(getDict().getUnicharset());
}
// Insert initial "pain points" to join the newly classified blob
// with its left and right neighbors.
if (classified != NULL && !classified->empty()) {
if (pain_point.col > 0) {
pain_points->GeneratePainPoint(
pain_point.col - 1, pain_point.row, LM_PPTYPE_SHAPE, 0.0,
true, segsearch_max_char_wh_ratio, word_res);
}
if (pain_point.row + 1 < ratings->dimension()) {
pain_points->GeneratePainPoint(
pain_point.col, pain_point.row + 1, LM_PPTYPE_SHAPE, 0.0,
true, segsearch_max_char_wh_ratio, word_res);
}
}
(*pending)[pain_point.col].SetBlobClassified(pain_point.row);
}
/**
* @name chop_word_main
*
* Classify the blobs in this word and permute the results. Find the
* worst blob in the word and chop it up. Continue this process until
* a good answer has been found or all the blobs have been chopped up
* enough. The results are returned in the WERD_RES.
*/
void Wordrec::chop_word_main(WERD_RES *word) {
int num_blobs = word->chopped_word->NumBlobs();
if (word->ratings == NULL) {
word->ratings = new MATRIX(num_blobs, wordrec_max_join_chunks);
}
if (word->ratings->get(0, 0) == NULL) {
// Run initial classification.
for (int b = 0; b < num_blobs; ++b) {
BLOB_CHOICE_LIST* choices = classify_piece(word->seam_array, b, b,
"Initial:", word->chopped_word,
word->blamer_bundle);
word->ratings->put(b, b, choices);
}
} else {
// Blobs have been pre-classified. Set matrix cell for all blob choices
for (int col = 0; col < word->ratings->dimension(); ++col) {
for (int row = col; row < word->ratings->dimension() &&
row < col + word->ratings->bandwidth(); ++row) {
BLOB_CHOICE_LIST* choices = word->ratings->get(col, row);
if (choices != NULL) {
BLOB_CHOICE_IT bc_it(choices);
for (bc_it.mark_cycle_pt(); !bc_it.cycled_list(); bc_it.forward()) {
bc_it.data()->set_matrix_cell(col, row);
}
}
}
}
}
// Run Segmentation Search.
BestChoiceBundle best_choice_bundle(word->ratings->dimension());
SegSearch(word, &best_choice_bundle, word->blamer_bundle);
if (word->best_choice == NULL) {
// SegSearch found no valid paths, so just use the leading diagonal.
word->FakeWordFromRatings();
}
word->RebuildBestState();
// If we finished without a hyphen at the end of the word, let the next word
// be found in the dictionary.
if (word->word->flag(W_EOL) &&
!getDict().has_hyphen_end(*word->best_choice)) {
getDict().reset_hyphen_vars(true);
}
if (word->blamer_bundle != NULL && this->fill_lattice_ != NULL) {
CallFillLattice(*word->ratings, word->best_choices,
*word->uch_set, word->blamer_bundle);
}
if (wordrec_debug_level > 0) {
tprintf("Final Ratings Matrix:\n");
word->ratings->print(getDict().getUnicharset());
}
word->FilterWordChoices(getDict().stopper_debug_level);
}
/// Resegments the word to achieve the target_text from the classifier.
/// Returns false if the re-segmentation fails.
/// Uses brute-force combination of up to #kMaxGroupSize adjacent blobs, and
/// applies a full search on the classifier results to find the best classified
/// segmentation. As a compromise to obtain better recall, 1-1 ambiguity
/// substitutions ARE used.
bool Tesseract::FindSegmentation(const GenericVector<UNICHAR_ID>& target_text,
WERD_RES* word_res) {
// Classify all required combinations of blobs and save results in choices.
int word_length = word_res->box_word->length();
GenericVector<BLOB_CHOICE_LIST*>* choices =
new GenericVector<BLOB_CHOICE_LIST*>[word_length];
for (int i = 0; i < word_length; ++i) {
for (int j = 1; j <= kMaxGroupSize && i + j <= word_length; ++j) {
BLOB_CHOICE_LIST* match_result = classify_piece(
word_res->seam_array, i, i + j - 1, "Applybox",
word_res->chopped_word, word_res->blamer_bundle);
if (applybox_debug > 2) {
tprintf("%d+%d:", i, j);
print_ratings_list("Segment:", match_result, unicharset);
}
choices[i].push_back(match_result);
}
}
// Search the segmentation graph for the target text. Must be an exact
// match. Using wildcards makes it difficult to find the correct
// segmentation even when it is there.
word_res->best_state.clear();
GenericVector<int> search_segmentation;
float best_rating = 0.0f;
SearchForText(choices, 0, word_length, target_text, 0, 0.0f,
&search_segmentation, &best_rating, &word_res->best_state);
for (int i = 0; i < word_length; ++i)
choices[i].delete_data_pointers();
delete [] choices;
if (word_res->best_state.empty()) {
// Build the original segmentation and if it is the same length as the
// truth, assume it will do.
int blob_count = 1;
for (int s = 0; s < word_res->seam_array.size(); ++s) {
SEAM* seam = word_res->seam_array[s];
if (!seam->HasAnySplits()) {
word_res->best_state.push_back(blob_count);
blob_count = 1;
} else {
++blob_count;
}
}
word_res->best_state.push_back(blob_count);
if (word_res->best_state.size() != target_text.size()) {
word_res->best_state.clear(); // No good. Original segmentation bad size.
return false;
}
}
word_res->correct_text.clear();
for (int i = 0; i < target_text.size(); ++i) {
word_res->correct_text.push_back(
STRING(unicharset.id_to_unichar(target_text[i])));
}
return true;
}
/**********************************************************************
* get_piece_rating
*
* Check to see if this piece has already been classified. If it has
* return that rating. Otherwise build the piece from the smaller
* pieces, classify it, store the rating for later, and take the piece
* apart again.
**********************************************************************/
BLOB_CHOICE_LIST *Wordrec::get_piece_rating(MATRIX *ratings,
TBLOB *blobs,
SEAMS seams,
inT16 start,
inT16 end) {
BLOB_CHOICE_LIST *choices = ratings->get(start, end);
if (choices == NOT_CLASSIFIED) {
choices = classify_piece(blobs,
seams,
start,
end);
ratings->put(start, end, choices);
}
return (choices);
}
/**********************************************************************
* get_piece_rating
*
* Check to see if this piece has already been classified. If it has
* return that rating. Otherwise build the piece from the smaller
* pieces, classify it, store the rating for later, and take the piece
* apart again.
**********************************************************************/
BLOB_CHOICE_LIST *Wordrec::get_piece_rating(MATRIX *ratings,
TBLOB *blobs,
const DENORM& denorm,
SEAMS seams,
inT16 start,
inT16 end,
BlamerBundle *blamer_bundle) {
BLOB_CHOICE_LIST *choices = ratings->get(start, end);
if (choices == NOT_CLASSIFIED) {
choices = classify_piece(blobs,
denorm,
seams,
start,
end,
blamer_bundle);
ratings->put(start, end, choices);
if (wordrec_debug_level > 1) {
tprintf("get_piece_rating(): updated ratings matrix\n");
ratings->print(getDict().getUnicharset());
}
}
return (choices);
}
void Wordrec::SegSearch(CHUNKS_RECORD *chunks_record,
WERD_CHOICE *best_choice,
BLOB_CHOICE_LIST_VECTOR *best_char_choices,
WERD_CHOICE *raw_choice,
STATE *output_best_state) {
int row, col = 0;
if (segsearch_debug_level > 0) {
tprintf("Starting SegSearch on ratings matrix:\n");
chunks_record->ratings->print(getDict().getUnicharset());
}
// Start with a fresh best_choice since rating adjustments
// used by the chopper and the new segmentation search are not compatible.
best_choice->set_rating(WERD_CHOICE::kBadRating);
// Clear best choice accumulator (that is used for adaption), so that
// choices adjusted by chopper do not interfere with the results from the
// segmentation search.
getDict().ClearBestChoiceAccum();
MATRIX *ratings = chunks_record->ratings;
// Priority queue containing pain points generated by the language model
// The priority is set by the language model components, adjustments like
// seam cost and width priority are factored into the priority.
HEAP *pain_points = MakeHeap(segsearch_max_pain_points);
// best_path_by_column records the lowest cost path found so far for each
// column of the chunks_record->ratings matrix over all the rows.
BestPathByColumn *best_path_by_column =
new BestPathByColumn[ratings->dimension()];
for (col = 0; col < ratings->dimension(); ++col) {
best_path_by_column[col].avg_cost = WERD_CHOICE::kBadRating;
best_path_by_column[col].best_vse = NULL;
}
language_model_->InitForWord(prev_word_best_choice_, &denorm_,
assume_fixed_pitch_char_segment,
best_choice->certainty(),
segsearch_max_char_wh_ratio,
pain_points, chunks_record);
MATRIX_COORD *pain_point;
float pain_point_priority;
BestChoiceBundle best_choice_bundle(
output_best_state, best_choice, raw_choice, best_char_choices);
// pending[i] stores a list of the parent/child pair of BLOB_CHOICE_LISTs,
// where i is the column of the child. Initially all the classified entries
// in the ratings matrix from column 0 (with parent NULL) are inserted into
// pending[0]. As the language model state is updated, new child/parent
// pairs are inserted into the lists. Next, the entries in pending[1] are
// considered, and so on. It is important that during the update the
// children are considered in the non-decreasing order of their column, since
// this guarantess that all the parents would be up to date before an update
// of a child is done.
SEG_SEARCH_PENDING_LIST *pending =
new SEG_SEARCH_PENDING_LIST[ratings->dimension()];
// Search for the ratings matrix for the initial best path.
for (row = 0; row < ratings->dimension(); ++row) {
if (ratings->get(0, row) != NOT_CLASSIFIED) {
pending[0].add_sorted(
SEG_SEARCH_PENDING::compare, true,
new SEG_SEARCH_PENDING(row, NULL, LanguageModel::kAllChangedFlag));
}
}
UpdateSegSearchNodes(0, &pending, &best_path_by_column, chunks_record,
pain_points, &best_choice_bundle);
// Keep trying to find a better path by fixing the "pain points".
int num_futile_classifications = 0;
while (!(language_model_->AcceptableChoiceFound() ||
num_futile_classifications >=
segsearch_max_futile_classifications)) {
// Get the next valid "pain point".
int pop;
while (true) {
pop = HeapPop(pain_points, &pain_point_priority, &pain_point);
if (pop == EMPTY) break;
if (pain_point->Valid(*ratings) &&
ratings->get(pain_point->col, pain_point->row) == NOT_CLASSIFIED) {
break;
} else {
delete pain_point;
}
}
if (pop == EMPTY) {
if (segsearch_debug_level > 0) tprintf("Pain points queue is empty\n");
break;
}
if (segsearch_debug_level > 0) {
tprintf("Classifying pain point priority=%.4f, col=%d, row=%d\n",
pain_point_priority, pain_point->col, pain_point->row);
}
BLOB_CHOICE_LIST *classified = classify_piece(
chunks_record->chunks, chunks_record->splits,
pain_point->col, pain_point->row);
ratings->put(pain_point->col, pain_point->row, classified);
if (segsearch_debug_level > 0) {
print_ratings_list("Updated ratings matrix with a new entry:",
ratings->get(pain_point->col, pain_point->row),
getDict().getUnicharset());
chunks_record->ratings->print(getDict().getUnicharset());
}
// Insert initial "pain points" to join the newly classified blob
// with its left and right neighbors.
if (!classified->empty()) {
float worst_piece_cert;
bool fragmented;
if (pain_point->col > 0) {
language_model_->GetWorstPieceCertainty(
pain_point->col-1, pain_point->row, chunks_record->ratings,
&worst_piece_cert, &fragmented);
language_model_->GeneratePainPoint(
pain_point->col-1, pain_point->row, false,
LanguageModel::kInitialPainPointPriorityAdjustment,
worst_piece_cert, fragmented, best_choice->certainty(),
segsearch_max_char_wh_ratio, NULL, NULL,
chunks_record, pain_points);
}
if (pain_point->row+1 < ratings->dimension()) {
language_model_->GetWorstPieceCertainty(
pain_point->col, pain_point->row+1, chunks_record->ratings,
&worst_piece_cert, &fragmented);
language_model_->GeneratePainPoint(
pain_point->col, pain_point->row+1, true,
LanguageModel::kInitialPainPointPriorityAdjustment,
worst_piece_cert, fragmented, best_choice->certainty(),
segsearch_max_char_wh_ratio, NULL, NULL,
chunks_record, pain_points);
}
}
// Record a pending entry with the pain_point and each of its parents.
int parent_row = pain_point->col - 1;
if (parent_row < 0) { // this node has no parents
pending[pain_point->col].add_sorted(
SEG_SEARCH_PENDING::compare, true,
new SEG_SEARCH_PENDING(pain_point->row, NULL,
LanguageModel::kAllChangedFlag));
} else {
for (int parent_col = 0; parent_col < pain_point->col; ++parent_col) {
if (ratings->get(parent_col, parent_row) != NOT_CLASSIFIED) {
pending[pain_point->col].add_sorted(
SEG_SEARCH_PENDING::compare, true,
new SEG_SEARCH_PENDING(pain_point->row,
ratings->get(parent_col, parent_row),
LanguageModel::kAllChangedFlag));
}
}
}
UpdateSegSearchNodes(pain_point->col, &pending, &best_path_by_column,
chunks_record, pain_points, &best_choice_bundle);
if (!best_choice_bundle.updated) ++num_futile_classifications;
if (segsearch_debug_level > 0) {
tprintf("num_futile_classifications %d\n", num_futile_classifications);
}
// Clean up
best_choice_bundle.updated = false;
delete pain_point; // done using this pain point
}
if (segsearch_debug_level > 0) {
tprintf("Done with SegSearch (AcceptableChoiceFound: %d\n",
language_model_->AcceptableChoiceFound());
}
// Clean up.
FreeHeapData(pain_points, MATRIX_COORD::Delete);
delete[] best_path_by_column;
delete[] pending;
for (row = 0; row < ratings->dimension(); ++row) {
for (col = 0; col <= row; ++col) {
BLOB_CHOICE_LIST *rating = ratings->get(col, row);
if (rating != NOT_CLASSIFIED) language_model_->DeleteState(rating);
}
}
}
