/*! * recogCreateFromPixa() * * Input: pixa (of labelled, 1 bpp images) * scalew (scale all widths to this; use 0 for no scaling) * scaleh (scale all heights to this; use 0 for no scaling) * templ_type (L_USE_AVERAGE or L_USE_ALL) * threshold (for binarization; typically ~128) * maxyshift (from nominal centroid alignment; typically 0 or 1) * fontdir (<optional> directory for bitmap fonts for debugging) * Return: recog, or null on error * * Notes: * (1) This is a convenience function for training from labelled data. * The pixa can be read from file. * (2) The pixa should contain the unscaled bitmaps used for training. * (3) The characters here should work as a single "font", because * each image example is put into a class defined by its * character label. All examples in the same class should be * similar. */ L_RECOG * recogCreateFromPixa(PIXA *pixa, l_int32 scalew, l_int32 scaleh, l_int32 templ_type, l_int32 threshold, l_int32 maxyshift, const char *fontdir) { char *text; l_int32 full, n, i, ntext; L_RECOG *recog; PIX *pix; PROCNAME("recogCreateFromPixa"); if (!pixa) return (L_RECOG *)ERROR_PTR("pixa not defined", procName, NULL); if (pixaVerifyDepth(pixa, NULL) != 1) return (L_RECOG *)ERROR_PTR("not all pix are 1 bpp", procName, NULL); pixaIsFull(pixa, &full, NULL); if (!full) return (L_RECOG *)ERROR_PTR("not all pix are present", procName, NULL); n = pixaGetCount(pixa); pixaCountText(pixa, &ntext); if (ntext == 0) return (L_RECOG *)ERROR_PTR("no pix have text strings", procName, NULL); if (ntext < n) L_ERROR("%d text strings < %d pix\n", procName, ntext, n); recog = recogCreate(scalew, scaleh, templ_type, threshold, maxyshift, fontdir); if (!recog) return (L_RECOG *)ERROR_PTR("recog not made", procName, NULL); for (i = 0; i < n; i++) { pix = pixaGetPix(pixa, i, L_CLONE); text = pixGetText(pix); if (!text || strlen(text) == 0) { L_ERROR("pix[%d] has no text\n", procName, i); pixDestroy(&pix); continue; } recogTrainLabelled(recog, pix, NULL, text, 0, 0); pixDestroy(&pix); } recogTrainingFinished(recog, 0); return recog; }
int main(int argc, char **argv) { char buf[32]; char *filein, *fileout, *fontdir, *textstr; l_int32 n, i, maxdepth, ntext, border, lossless, display, showtext; l_float32 scalefact; L_BMF *bmf; PIX *pix1, *pix2, *pix3, *pix4, *pixd; PIXA *pixa, *pixad; static char mainName[] = "displaypixa"; if (argc != 3 && argc != 4 && argc != 7 && argc != 8) { fprintf(stderr, "Syntax error in displaypixa:\n" " displaypixa filein fileout [showtext]\n" " displaypixa filein scalefact border" " lossless disp fileout [showtext]\n"); return 1; } filein = argv[1]; if ((pixa = pixaRead(filein)) == NULL) return ERROR_INT("pixa not made", mainName, 1); pixaCountText(pixa, &ntext); if (argc == 3 || argc == 4) fileout = argv[2]; if (argc == 4) showtext = atoi(argv[3]); /* Simple specification; no output text */ if (argc == 3 || (argc == 4 && (ntext == 0 || showtext == 0))) { /* no text output */ pixaVerifyDepth(pixa, &maxdepth); pixd = pixaDisplayTiledInRows(pixa, maxdepth, 1400, 1.0, 0, 10, 0); pixDisplay(pixd, 100, 100); if (pixGetDepth(pixd) == 1) pixWrite(fileout, pixd, IFF_PNG); else pixWrite(fileout, pixd, IFF_JFIF_JPEG); pixDestroy(&pixd); pixaDestroy(&pixa); return 0; } /* Simple specification with output text */ if (argc == 4) { /* showtext == 1 && ntext > 0 */ n = pixaGetCount(pixa); bmf = bmfCreate(NULL, 6); pixad = pixaCreate(n); for (i = 0; i < n; i++) { pix1 = pixaGetPix(pixa, i, L_CLONE); pix2 = pixConvertTo32(pix1); pix3 = pixAddBorderGeneral(pix2, 10, 10, 5, 5, 0xffffff00); textstr = pixGetText(pix1); if (textstr && strlen(textstr) > 0) { snprintf(buf, sizeof(buf), "%s", textstr); pix4 = pixAddSingleTextblock(pix3, bmf, buf, 0xff000000, L_ADD_BELOW, NULL); } else { pix4 = pixClone(pix3); } pixaAddPix(pixad, pix4, L_INSERT); pixDestroy(&pix1); pixDestroy(&pix2); pixDestroy(&pix3); } bmfDestroy(&bmf); pixaVerifyDepth(pixad, &maxdepth); pixd = pixaDisplayTiledInRows(pixad, maxdepth, 1400, 1.0, 0, 10, 0); pixDisplay(pixd, 100, 100); if (pixGetDepth(pixd) == 1) pixWrite(fileout, pixd, IFF_PNG); else pixWrite(fileout, pixd, IFF_JFIF_JPEG); pixDestroy(&pixd); pixaDestroy(&pixa); pixaDestroy(&pixad); return 0; } /* Full specification */ scalefact = atof(argv[2]); border = atoi(argv[3]); lossless = atoi(argv[4]); display = atoi(argv[5]); fileout = argv[6]; showtext = (argc == 8) ? atoi(argv[7]) : 0; if (showtext && ntext == 0) L_INFO("No text found in any of the pix\n", mainName); bmf = (showtext && ntext > 0) ? bmfCreate(NULL, 6) : NULL; n = pixaGetCount(pixa); pixad = pixaCreate(n); for (i = 0; i < n; i++) { pix1 = pixaGetPix(pixa, i, L_CLONE); pix2 = pixConvertTo32(pix1); pix3 = pixAddBorderGeneral(pix2, 10, 10, 5, 5, 0xffffff00); textstr = pixGetText(pix1); if (bmf && textstr && strlen(textstr) > 0) { snprintf(buf, sizeof(buf), "%s", textstr); pix4 = pixAddSingleTextblock(pix3, bmf, buf, 0xff000000, L_ADD_BELOW, NULL); } else { pix4 = pixClone(pix3); } pixaAddPix(pixad, pix4, L_INSERT); pixDestroy(&pix1); pixDestroy(&pix2); pixDestroy(&pix3); } bmfDestroy(&bmf); pixaVerifyDepth(pixad, &maxdepth); pixd = pixaDisplayTiledInRows(pixad, maxdepth, 1400, scalefact, 0, 10, border); if (display) pixDisplay(pixd, 20, 20); if (pixGetDepth(pixd) == 1 || lossless) pixWrite(fileout, pixd, IFF_PNG); else pixWrite(fileout, pixd, IFF_JFIF_JPEG); pixDestroy(&pixd); pixaDestroy(&pixa); pixaDestroy(&pixad); return 0; }