|
Tesseract
3.02
|
|
Tesseract
3.02
|
#include <word_size_model.h>
Public Member Functions | |
| WordSizeModel (CharSet *, bool contextual) | |
| virtual | ~WordSizeModel () |
| int | Cost (CharSamp **samp_array, int samp_cnt) const |
| bool | Save (string file_name) |
| int | FontCount () const |
| const FontPairSizeInfo * | FontInfo () const |
Static Public Member Functions | |
| static WordSizeModel * | Create (const string &data_file_path, const string &lang, CharSet *char_set, bool contextual) |
| static double | PairCost (int width_0, int height_0, int top_0, int width_1, int height_1, int top_1, const PairSizeInfo &pair_info) |
| static int | SizeCode (int cls_id, int start, int end) |
Definition at line 53 of file word_size_model.h.
| tesseract::WordSizeModel::WordSizeModel | ( | CharSet * | char_set, |
| bool | contextual | ||
| ) |
Definition at line 28 of file word_size_model.cpp.
{
char_set_ = char_set;
contextual_ = contextual;
}
| tesseract::WordSizeModel::~WordSizeModel | ( | ) | [virtual] |
Definition at line 33 of file word_size_model.cpp.
{
for (int fnt = 0; fnt < font_pair_size_models_.size(); fnt++) {
FontPairSizeInfo fnt_info = font_pair_size_models_[fnt];
delete []fnt_info.pair_size_info[0];
delete []fnt_info.pair_size_info;
}
}
| int tesseract::WordSizeModel::Cost | ( | CharSamp ** | samp_array, |
| int | samp_cnt | ||
| ) | const |
Definition at line 210 of file word_size_model.cpp.
{
if (samp_cnt < 2) {
return 0;
}
double best_dist = static_cast<double>(WORST_COST);
int best_fnt = -1;
for (int fnt = 0; fnt < font_pair_size_models_.size(); fnt++) {
const FontPairSizeInfo *fnt_info = &font_pair_size_models_[fnt];
double mean_dist = 0;
int pair_cnt = 0;
for (int smp_0 = 0; smp_0 < samp_cnt; smp_0++) {
int cls_0 = char_set_->ClassID(samp_array[smp_0]->StrLabel());
if (cls_0 < 1) {
continue;
}
// compute size code for samp 0 based on class id and position
int size_code_0;
if (contextual_) {
size_code_0 = SizeCode(cls_0,
samp_array[smp_0]->FirstChar() == 0 ? 0 : 1,
samp_array[smp_0]->LastChar() == 0 ? 0 : 1);
} else {
size_code_0 = cls_0;
}
int char0_height = samp_array[smp_0]->Height();
int char0_width = samp_array[smp_0]->Width();
int char0_top = samp_array[smp_0]->Top();
for (int smp_1 = smp_0 + 1; smp_1 < samp_cnt; smp_1++) {
int cls_1 = char_set_->ClassID(samp_array[smp_1]->StrLabel());
if (cls_1 < 1) {
continue;
}
// compute size code for samp 0 based on class id and position
int size_code_1;
if (contextual_) {
size_code_1 = SizeCode(cls_1,
samp_array[smp_1]->FirstChar() == 0 ? 0 : 1,
samp_array[smp_1]->LastChar() == 0 ? 0 : 1);
} else {
size_code_1 = cls_1;
}
double dist = PairCost(
char0_width, char0_height, char0_top, samp_array[smp_1]->Width(),
samp_array[smp_1]->Height(), samp_array[smp_1]->Top(),
fnt_info->pair_size_info[size_code_0][size_code_1]);
if (dist > 0) {
mean_dist += dist;
pair_cnt++;
}
} // smp_1
} // smp_0
if (pair_cnt == 0) {
continue;
}
mean_dist /= pair_cnt;
if (best_fnt == -1 || mean_dist < best_dist) {
best_dist = mean_dist;
best_fnt = fnt;
}
}
if (best_fnt == -1) {
return static_cast<int>(WORST_COST);
} else {
return static_cast<int>(best_dist);
}
}
| WordSizeModel * tesseract::WordSizeModel::Create | ( | const string & | data_file_path, |
| const string & | lang, | ||
| CharSet * | char_set, | ||
| bool | contextual | ||
| ) | [static] |
Definition at line 41 of file word_size_model.cpp.
{
WordSizeModel *obj = new WordSizeModel(char_set, contextual);
if (!obj) {
fprintf(stderr, "Cube ERROR (WordSizeModel::Create): unable to allocate "
"new word size model object\n");
return NULL;
}
if (!obj->Init(data_file_path, lang)) {
delete obj;
return NULL;
}
return obj;
}
| int tesseract::WordSizeModel::FontCount | ( | ) | const [inline] |
Definition at line 72 of file word_size_model.h.
{
return font_pair_size_models_.size();
}
| const FontPairSizeInfo* tesseract::WordSizeModel::FontInfo | ( | ) | const [inline] |
Definition at line 75 of file word_size_model.h.
{
return &font_pair_size_models_[0];
}
| double tesseract::WordSizeModel::PairCost | ( | int | width_0, |
| int | height_0, | ||
| int | top_0, | ||
| int | width_1, | ||
| int | height_1, | ||
| int | top_1, | ||
| const PairSizeInfo & | pair_info | ||
| ) | [static] |
Definition at line 280 of file word_size_model.cpp.
{
double scale_factor = static_cast<double>(pair_info.hgt_0) /
static_cast<double>(height_0);
double dist = 0.0;
if (scale_factor > 0) {
double norm_width_0 = width_0 * scale_factor;
double norm_width_1 = width_1 * scale_factor;
double norm_height_1 = height_1 * scale_factor;
double norm_delta_top = (top_1 - top_0) * scale_factor;
// accumulate the distance between the model character and the
// predicted one on all dimensions of the pair
dist += fabs(pair_info.wid_0 - norm_width_0);
dist += fabs(pair_info.wid_1 - norm_width_1);
dist += fabs(pair_info.hgt_1 - norm_height_1);
dist += fabs(pair_info.delta_top - norm_delta_top);
}
return dist;
}
| bool tesseract::WordSizeModel::Save | ( | string | file_name | ) |
| static int tesseract::WordSizeModel::SizeCode | ( | int | cls_id, |
| int | start, | ||
| int | end | ||
| ) | [inline, static] |
Definition at line 80 of file word_size_model.h.
{
return (cls_id << 2) + (end << 1) + start;
}
1.7.5.1