/*====================================================================* - Copyright (C) 2001 Leptonica. All rights reserved. - - Redistribution and use in source and binary forms, with or without - modification, are permitted provided that the following conditions - are met: - 1. Redistributions of source code must retain the above copyright - notice, this list of conditions and the following disclaimer. - 2. Redistributions in binary form must reproduce the above - copyright notice, this list of conditions and the following - disclaimer in the documentation and/or other materials - provided with the distribution. - - THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS - ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT - LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR - A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL ANY - CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, - EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, - PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR - PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY - OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING - NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS - SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. *====================================================================*/ /* * recogtest4.c * * Test document image decoding (DID) approach to splitting characters * This tests the low-level recogDecode() function. * Splitting succeeds for both with and without character height scaling. * * But cf. recogtest5.c. Note that recogIdentifyMultiple(), which * does prefiltering and splitting before character identification, * does not accept input that has been scaled. That is because * the only reason for scaling the templates is that the recognizer * is a hybrid BAR/BSR, where we've used a mixture of templates from * a single source and bootstrap templates from many sources. */ #ifdef HAVE_CONFIG_H #include #endif /* HAVE_CONFIG_H */ #include "string.h" #include "allheaders.h" static PIX *GetBigComponent(PIX *pixs); l_int32 main(int argc, char **argv) { char buf[256]; l_int32 i, item; l_int32 example[6] = {17, 20, 21, 22, 23, 24}; /* for decoding */ BOXA *boxa; PIX *pix1, *pix2, *pixdb; PIXA *pixa1, *pixa2; L_RECOG *recog; if (argc != 1) { lept_stderr(" Syntax: recogtest4\n"); return 1; } setLeptDebugOK(1); lept_mkdir("lept/recog"); /* Generate the recognizer */ pixa1 = pixaRead("recog/sets/train01.pa"); #if 1 /* scale to fixed height */ recog = recogCreateFromPixa(pixa1, 0, 40, 0, 128, 1); #else /* no scaling */ recog = recogCreateFromPixa(pixa1, 0, 0, 0, 128, 1); #endif recogAverageSamples(&recog, 1); recogWrite("/tmp/lept/recog/rec1.rec", recog); /* Show the templates */ recogDebugAverages(&recog, 1); if (!recog) { lept_stderr("Averaging failed!!\n"); return 1; } recogShowMatchesInRange(recog, recog->pixa_tr, 0.0, 1.0, 1); /* Get a set of problem images to decode */ pixa2 = pixaRead("recog/sets/test01.pa"); /* Decode a subset of them. It takes about 1 ms to decode a * 4 digit number, with both Viterbi and rescoring (debug off). */ for (i = 0; i < 6; i++) { /* if (i != 3) continue; */ /* remove this comment to do all 6 */ item = example[i]; pix1 = pixaGetPix(pixa2, item, L_CLONE); pixDisplay(pix1, 100, 100); pix2 = GetBigComponent(pix1); boxa = recogDecode(recog, pix2, 2, &pixdb); pixDisplay(pixdb, 300, 100); snprintf(buf, sizeof(buf), "/tmp/lept/recog/did-%d.png", item); pixWrite(buf, pixdb, IFF_PNG); pixDestroy(&pixdb); boxaDestroy(&boxa); pixDestroy(&pix1); pixDestroy(&pix2); } pixaDestroy(&pixa1); pixaDestroy(&pixa2); recogDestroy(&recog); return 0; } static PIX * GetBigComponent(PIX *pixs) { BOX *box; PIX *pix1, *pixd; pix1 = pixMorphSequence(pixs, "c40.7 + o20.15 + d25.1", 0); pixClipToForeground(pix1, NULL, &box); pixd = pixClipRectangle(pixs, box, NULL); pixDestroy(&pix1); boxDestroy(&box); return pixd; }