/*---------------------------------------------------------------------------*/
void
LearnBlob (TBLOB * Blob, TEXTROW * Row, char BlobText[], int TextLength)
/*
** Parameters:
** Blob blob whose micro-features are to be learned
** Row row of text that blob came from
** BlobText text that corresponds to blob
** TextLength number of characters in blob
** Globals:
** imagefile base filename of the page being learned
** FontName name of font currently being trained on
** Operation:
** Extract micro-features from the specified blob and append
** them to the appropriate file.
** Return: none
** Exceptions: none
** History: 7/28/89, DSJ, Created.
*/
#define MAXFILENAME 80
#define MAXCHARNAME 20
#define MAXFONTNAME 20
#define TRAIN_SUFFIX ".tr"
{
static FILE *FeatureFile = NULL;
char Filename[MAXFILENAME];
char CharName[MAXCHARNAME];
CHAR_DESC CharDesc;
LINE_STATS LineStats;
EnterLearnMode;
// throw out blobs which do not represent only one character
if (TextLength != 1)
return;
GetLineStatsFromRow(Row, &LineStats);
CharDesc = ExtractBlobFeatures (Blob, &LineStats);
// if a feature file is not yet open, open it
// the name of the file is the name of the image plus TRAIN_SUFFIX
if (FeatureFile == NULL) {
strcpy(Filename, imagefile);
strcat(Filename, TRAIN_SUFFIX);
FeatureFile = Efopen (Filename, "w");
cprintf ("TRAINING ... Font name = %s.\n", FontName);
}
// get the name of the character for this blob
chartoname (CharName, BlobText[0], "");
// label the features with a class name and font name
fprintf (FeatureFile, "\n%s %s ", FontName, CharName);
// write micro-features to file and clean up
WriteCharDescription(FeatureFile, CharDesc);
FreeCharDescription(CharDesc);
} // LearnBlob
void LearnBlob(FILE* FeatureFile, TBLOB* Blob, TEXTROW* Row,
const char* BlobText, const char* FontName) {
CHAR_DESC CharDesc;
LINE_STATS LineStats;
EnterLearnMode;
GetLineStatsFromRow(Row, &LineStats);
CharDesc = ExtractBlobFeatures (Blob, &LineStats);
if (CharDesc == NULL) {
cprintf("LearnBLob: CharDesc was NULL. Aborting.\n");
return;
}
// label the features with a class name and font name
fprintf (FeatureFile, "\n%s %s ", FontName, BlobText);
// write micro-features to file and clean up
WriteCharDescription(FeatureFile, CharDesc);
FreeCharDescription(CharDesc);
} // LearnBlob
// Adapt to recognize the current image as the given character.
// The image must be preloaded and be just an image of a single character.
void TessBaseAPI::AdaptToCharacter(const char *unichar_repr,
int length,
float baseline,
float xheight,
float descender,
float ascender) {
UNICHAR_ID id = unicharset.unichar_to_id(unichar_repr, length);
LINE_STATS LineStats;
TEXTROW row;
fill_dummy_row(baseline, xheight, descender, ascender, &row);
GetLineStatsFromRow(&row, &LineStats);
TBLOB *blob = make_tesseract_blob(baseline, xheight, descender, ascender);
float threshold;
int best_class = 0;
float best_rating = -100;
// Classify to get a raw choice.
LIST result = AdaptiveClassifier(blob, NULL, &row);
LIST p;
for (p = result; p != NULL; p = p->next) {
A_CHOICE *tesschoice = (A_CHOICE *) p->node;
if (tesschoice->rating > best_rating) {
best_rating = tesschoice->rating;
best_class = tesschoice->string[0];
}
}
FLOAT32 GetBestRatingFor(TBLOB *Blob, LINE_STATS *LineStats, CLASS_ID ClassId);
// We have to use char-level adaptation because otherwise
// someone should do forced alignment somewhere.
void AdaptToChar(TBLOB *Blob,
LINE_STATS *LineStats,
CLASS_ID ClassId,
FLOAT32 Threshold);
if (id == best_class)
threshold = GoodAdaptiveMatch;
else {
/* the blob was incorrectly classified - find the rating threshold
needed to create a template which will correct the error with
some margin. However, don't waste time trying to make
templates which are too tight. */
threshold = GetBestRatingFor(blob, &LineStats, id);
threshold *= .9;
const float max_threshold = .125;
const float min_threshold = .02;
if (threshold > max_threshold)
threshold = max_threshold;
// I have cuddled the following line to set it out of the strike
// of the coverage testing tool. I have no idea how to trigger
// this situation nor I have any necessity to do it. --mezhirov
if (threshold < min_threshold) threshold = min_threshold;
}
if (blob->outlines)
AdaptToChar(blob, &LineStats, id, threshold);
free_blob(blob);
}
/*---------------------------------------------------------------------------*/
void
LearnBlob (const STRING& filename,
TBLOB * Blob, TEXTROW * Row, char BlobText[])
/*
** Parameters:
** Blob blob whose micro-features are to be learned
** Row row of text that blob came from
** BlobText text that corresponds to blob
** TextLength number of characters in blob
** Globals:
** imagefile base filename of the page being learned
** classify_font_name
** name of font currently being trained on
** Operation:
** Extract micro-features from the specified blob and append
** them to the appropriate file.
** Return: none
** Exceptions: none
** History: 7/28/89, DSJ, Created.
*/
#define TRAIN_SUFFIX ".tr"
{
static FILE *FeatureFile = NULL;
STRING Filename(filename);
CHAR_DESC CharDesc;
LINE_STATS LineStats;
EnterLearnMode;
GetLineStatsFromRow(Row, &LineStats);
CharDesc = ExtractBlobFeatures (Blob, &LineStats);
if (CharDesc == NULL) {
cprintf("LearnBLob: CharDesc was NULL. Aborting.\n");
return;
}
// If no fontname was set, try to extract it from the filename
char CurrFontName[32] = "";
strncpy(CurrFontName, static_cast<STRING>(classify_font_name).string(), 32);
/*
if (!strcmp(CurrFontName, "UnknownFont")) {
// filename is expected to be of the form [lang].[fontname].exp[num]
// The [lang], [fontname] and [num] fields should not have '.' characters.
const char *basename = strrchr(filename.string(), '/');
const char *firstdot = strchr(basename, '.');
const char *lastdot = strrchr(filename.string(), '.');
if (firstdot != lastdot && firstdot != NULL && lastdot != NULL) {
strncpy(CurrFontName, firstdot + 1, lastdot - firstdot - 1);
}
}
//*/
// if a feature file is not yet open, open it
// the name of the file is the name of the image plus TRAIN_SUFFIX
if (FeatureFile == NULL) {
Filename += TRAIN_SUFFIX;
FeatureFile = Efopen (Filename.string(), "w");
cprintf ("TRAINING ... Font name = %s\n", CurrFontName);
}
// label the features with a class name and font name
fprintf (FeatureFile, "\n%s %s ", CurrFontName, BlobText);
// write micro-features to file and clean up
WriteCharDescription(FeatureFile, CharDesc);
FreeCharDescription(CharDesc);
} // LearnBlob
