/*!
* pixaRemovePix()
*
* Input: pixa
* index (of pix to be removed)
* Return: 0 if OK, 1 on error
*
* Notes:
* (1) This shifts pixa[i] --> pixa[i - 1] for all i > index.
* (2) It should not be used repeatedly on large arrays,
* because the function is O(n).
* (3) The corresponding box is removed as well, if it exists.
*/
l_int32
pixaRemovePix(PIXA *pixa,
l_int32 index)
{
l_int32 i, n, nbox;
BOXA *boxa;
PIX **array;
PROCNAME("pixaRemovePix");
if (!pixa)
return ERROR_INT("pixa not defined", procName, 1);
n = pixaGetCount(pixa);
if (index < 0 || index >= n)
return ERROR_INT("index not in {0...n - 1}", procName, 1);
/* Remove the pix */
array = pixa->pix;
pixDestroy(&array[index]);
for (i = index + 1; i < n; i++)
array[i - 1] = array[i];
array[n - 1] = NULL;
pixa->n--;
/* Remove the box if it exists */
boxa = pixa->boxa;
nbox = boxaGetCount(boxa);
if (index < nbox)
boxaRemoveBox(boxa, index);
return 0;
}
/*!
* pixaGenerateFont()
*
* Input: pix (of 95 characters in 3 rows)
* fontsize (4, 6, 8, ... , 20, in pts at 300 ppi)
* &bl1 (<return> baseline of row 1)
* &bl2 (<return> baseline of row 2)
* &bl3 (<return> baseline of row 3)
* Return: pixa of font bitmaps for 95 characters, or null on error
*
* Notes:
* (1) This does all the work. See pixaGenerateFontFromFile()
* for an overview.
* (2) The pix is for one of the 9 fonts. @fontsize is only
* used here for debugging.
*/
PIXA *
pixaGenerateFont(PIX *pixs,
l_int32 fontsize,
l_int32 *pbl0,
l_int32 *pbl1,
l_int32 *pbl2)
{
l_int32 i, j, nrows, nrowchars, nchars, h, yval;
l_int32 width, height;
l_int32 baseline[3];
l_int32 *tab = NULL;
BOX *box, *box1, *box2;
BOXA *boxar, *boxac, *boxacs;
PIX *pix1, *pix2, *pixr, *pixrc, *pixc;
PIXA *pixa;
l_int32 n, w, inrow, top;
l_int32 *ia;
NUMA *na;
PROCNAME("pixaGenerateFont");
if (!pbl0 || !pbl1 || !pbl2)
return (PIXA *)ERROR_PTR("&bl not all defined", procName, NULL);
*pbl0 = *pbl1 = *pbl2 = 0;
if (!pixs)
return (PIXA *)ERROR_PTR("pixs not defined", procName, NULL);
/* Locate the 3 rows of characters */
w = pixGetWidth(pixs);
na = pixCountPixelsByRow(pixs, NULL);
boxar = boxaCreate(0);
n = numaGetCount(na);
ia = numaGetIArray(na);
inrow = 0;
for (i = 0; i < n; i++) {
if (!inrow && ia[i] > 0) {
inrow = 1;
top = i;
} else if (inrow && ia[i] == 0) {
inrow = 0;
box = boxCreate(0, top, w, i - top);
boxaAddBox(boxar, box, L_INSERT);
}
}
FREE(ia);
numaDestroy(&na);
nrows = boxaGetCount(boxar);
#if DEBUG_FONT_GEN
L_INFO("For fontsize %s, have %d rows\n", procName, fontsize, nrows);
#endif /* DEBUG_FONT_GEN */
if (nrows != 3) {
L_INFO("nrows = %d; skipping fontsize %d\n", procName, nrows, fontsize);
return (PIXA *)ERROR_PTR("3 rows not generated", procName, NULL);
}
/* Grab the character images and baseline data */
#if DEBUG_BASELINE
lept_rmdir("baseline");
lept_mkdir("baseline");
#endif /* DEBUG_BASELINE */
tab = makePixelSumTab8();
pixa = pixaCreate(95);
for (i = 0; i < nrows; i++) {
box = boxaGetBox(boxar, i, L_CLONE);
pixr = pixClipRectangle(pixs, box, NULL); /* row of chars */
pixGetTextBaseline(pixr, tab, &yval);
baseline[i] = yval;
#if DEBUG_BASELINE
L_INFO("Baseline info: row %d, yval = %d, h = %d\n", procName,
i, yval, pixGetHeight(pixr));
pix1 = pixCopy(NULL, pixr);
pixRenderLine(pix1, 0, yval, pixGetWidth(pix1), yval, 1,
L_FLIP_PIXELS);
if (i == 0 )
pixWrite("/tmp/baseline/row0.png", pix1, IFF_PNG);
else if (i == 1)
pixWrite("/tmp/baseline/row1.png", pix1, IFF_PNG);
else
pixWrite("/tmp/baseline/row2.png", pix1, IFF_PNG);
pixDestroy(&pix1);
#endif /* DEBUG_BASELINE */
boxDestroy(&box);
pixrc = pixCloseSafeBrick(NULL, pixr, 1, 35);
boxac = pixConnComp(pixrc, NULL, 8);
boxacs = boxaSort(boxac, L_SORT_BY_X, L_SORT_INCREASING, NULL);
if (i == 0) { /* consolidate the two components of '"' */
box1 = boxaGetBox(boxacs, 1, L_CLONE);
box2 = boxaGetBox(boxacs, 2, L_CLONE);
box1->w = box2->x + box2->w - box1->x; /* increase width */
boxDestroy(&box1);
boxDestroy(&box2);
boxaRemoveBox(boxacs, 2);
}
h = pixGetHeight(pixr);
nrowchars = boxaGetCount(boxacs);
for (j = 0; j < nrowchars; j++) {
box = boxaGetBox(boxacs, j, L_COPY);
if (box->w <= 2 && box->h == 1) { /* skip 1x1, 2x1 components */
boxDestroy(&box);
continue;
}
box->y = 0;
box->h = h - 1;
pixc = pixClipRectangle(pixr, box, NULL);
boxDestroy(&box);
if (i == 0 && j == 0) /* add a pix for the space; change later */
pixaAddPix(pixa, pixc, L_COPY);
if (i == 2 && j == 0) /* add a pix for the '\'; change later */
pixaAddPix(pixa, pixc, L_COPY);
pixaAddPix(pixa, pixc, L_INSERT);
}
pixDestroy(&pixr);
pixDestroy(&pixrc);
boxaDestroy(&boxac);
boxaDestroy(&boxacs);
}
FREE(tab);
nchars = pixaGetCount(pixa);
if (nchars != 95)
return (PIXA *)ERROR_PTR("95 chars not generated", procName, NULL);
*pbl0 = baseline[0];
*pbl1 = baseline[1];
*pbl2 = baseline[2];
/* Fix the space character up; it should have no ON pixels,
* and be about twice as wide as the '!' character. */
pix1 = pixaGetPix(pixa, 0, L_CLONE);
width = 2 * pixGetWidth(pix1);
height = pixGetHeight(pix1);
pixDestroy(&pix1);
pix1 = pixCreate(width, height, 1);
pixaReplacePix(pixa, 0, pix1, NULL);
/* Fix up the '\' character; use a LR flip of the '/' char */
pix1 = pixaGetPix(pixa, 15, L_CLONE);
pix2 = pixFlipLR(NULL, pix1);
pixDestroy(&pix1);
pixaReplacePix(pixa, 60, pix2, NULL);
#if DEBUG_CHARS
pix1 = pixaDisplayTiled(pixa, 1500, 0, 10);
pixDisplay(pix1, 100 * i, 200);
pixDestroy(&pix1);
#endif /* DEBUG_CHARS */
boxaDestroy(&boxar);
return pixa;
}
/*!
* pixaGenerateFont()
*
* Input: dir (directory holding image of character set)
* size (4, 6, 8, ... , 20, in pts at 300 ppi)
* &bl1 (<return> baseline of row 1)
* &bl2 (<return> baseline of row 2)
* &bl3 (<return> baseline of row 3)
* Return: pixa of font bitmaps for 95 characters, or null on error
*
* These font generation functions use 9 sets, each with bitmaps
* of 94 ascii characters, all in Palatino-Roman font.
* Each input bitmap has 3 rows of characters. The range of
* ascii values in each row is as follows:
* row 0: 32-57 (32 is a space)
* row 1: 58-91 (92, '\', is not represented in this font)
* row 2: 93-126
* We LR flip the '/' char to generate a bitmap for the missing
* '\' character, so that we have representations of all 95
* printable chars.
*
* Computation of the bitmaps and baselines for a single
* font takes from 40 to 200 msec on a 2 GHz processor,
* depending on the size. Use pixaGetFont() to read the
* generated character set directly from files that were
* produced in prog/genfonts.c using this function.
*/
PIXA *
pixaGenerateFont(const char *dir,
l_int32 size,
l_int32 *pbl0,
l_int32 *pbl1,
l_int32 *pbl2)
{
char *pathname;
l_int32 fileno;
l_int32 i, j, nrows, nrowchars, nchars, h, yval;
l_int32 width, height;
l_int32 baseline[3];
l_int32 *tab;
BOX *box, *box1, *box2;
BOXA *boxar, *boxac, *boxacs;
PIX *pixs, *pixt1, *pixt2, *pixt3;
PIX *pixr, *pixrc, *pixc;
PIXA *pixa;
PROCNAME("pixaGenerateFont");
if (!pbl0 || !pbl1 || !pbl2)
return (PIXA *)ERROR_PTR("&bl not all defined", procName, NULL);
*pbl0 = *pbl1 = *pbl2 = 0;
fileno = (size / 2) - 2;
if (fileno < 0 || fileno > NFONTS)
return (PIXA *)ERROR_PTR("font size invalid", procName, NULL);
tab = makePixelSumTab8();
pathname = genPathname(dir, inputfonts[fileno]);
if ((pixs = pixRead(pathname)) == NULL)
return (PIXA *)ERROR_PTR("pixs not all defined", procName, NULL);
FREE(pathname);
pixa = pixaCreate(95);
pixt1 = pixMorphSequence(pixs, "c1.35 + c101.1", 0);
boxar = pixConnComp(pixt1, NULL, 8); /* one box for each row */
pixDestroy(&pixt1);
nrows = boxaGetCount(boxar);
#if DEBUG_FONT_GEN
fprintf(stderr, "For font %s, number of rows is %d\n",
inputfonts[fileno], nrows);
#endif /* DEBUG_FONT_GEN */
if (nrows != 3) {
L_INFO_INT2("nrows = %d; skipping font %d", procName, nrows, fileno);
return (PIXA *)ERROR_PTR("3 rows not generated", procName, NULL);
}
for (i = 0; i < nrows; i++) {
box = boxaGetBox(boxar, i, L_CLONE);
pixr = pixClipRectangle(pixs, box, NULL); /* row of chars */
pixGetTextBaseline(pixr, tab, &yval);
baseline[i] = yval;
#if DEBUG_BASELINE
{ PIX *pixbl;
fprintf(stderr, "row %d, yval = %d, h = %d\n",
i, yval, pixGetHeight(pixr));
pixbl = pixCopy(NULL, pixr);
pixRenderLine(pixbl, 0, yval, pixGetWidth(pixbl), yval, 1,
L_FLIP_PIXELS);
if (i == 0 )
pixWrite("junktl0", pixbl, IFF_PNG);
else if (i == 1)
pixWrite("junktl1", pixbl, IFF_PNG);
else
pixWrite("junktl2", pixbl, IFF_PNG);
pixDestroy(&pixbl);
}
#endif /* DEBUG_BASELINE */
boxDestroy(&box);
pixrc = pixCloseSafeBrick(NULL, pixr, 1, 35);
boxac = pixConnComp(pixrc, NULL, 8);
boxacs = boxaSort(boxac, L_SORT_BY_X, L_SORT_INCREASING, NULL);
if (i == 0) { /* consolidate the two components of '"' */
box1 = boxaGetBox(boxacs, 1, L_CLONE);
box2 = boxaGetBox(boxacs, 2, L_CLONE);
box1->w = box2->x + box2->w - box1->x; /* increase width */
boxDestroy(&box1);
boxDestroy(&box2);
boxaRemoveBox(boxacs, 2);
}
h = pixGetHeight(pixr);
nrowchars = boxaGetCount(boxacs);
for (j = 0; j < nrowchars; j++) {
box = boxaGetBox(boxacs, j, L_COPY);
if (box->w <= 2 && box->h == 1) { /* skip 1x1, 2x1 components */
boxDestroy(&box);
continue;
}
box->y = 0;
box->h = h - 1;
pixc = pixClipRectangle(pixr, box, NULL);
boxDestroy(&box);
if (i == 0 && j == 0) /* add a pix for the space; change later */
pixaAddPix(pixa, pixc, L_COPY);
if (i == 2 && j == 0) /* add a pix for the '\'; change later */
pixaAddPix(pixa, pixc, L_COPY);
pixaAddPix(pixa, pixc, L_INSERT);
}
pixDestroy(&pixr);
pixDestroy(&pixrc);
boxaDestroy(&boxac);
boxaDestroy(&boxacs);
}
nchars = pixaGetCount(pixa);
if (nchars != 95)
return (PIXA *)ERROR_PTR("95 chars not generated", procName, NULL);
*pbl0 = baseline[0];
*pbl1 = baseline[1];
*pbl2 = baseline[2];
/* Fix the space character up; it should have no ON pixels,
* and be about twice as wide as the '!' character. */
pixt2 = pixaGetPix(pixa, 0, L_CLONE);
width = 2 * pixGetWidth(pixt2);
height = pixGetHeight(pixt2);
pixDestroy(&pixt2);
pixt2 = pixCreate(width, height, 1);
pixaReplacePix(pixa, 0, pixt2, NULL);
/* Fix up the '\' character; use a LR flip of the '/' char */
pixt2 = pixaGetPix(pixa, 15, L_CLONE);
pixt3 = pixFlipLR(NULL, pixt2);
pixDestroy(&pixt2);
pixaReplacePix(pixa, 60, pixt3, NULL);
#if DEBUG_CHARS
{ PIX *pixd;
pixd = pixaDisplayTiled(pixa, 1500, 0, 10);
pixDisplay(pixd, 100 * i, 200);
pixDestroy(&pixd);
}
#endif /* DEBUG_CHARS */
pixDestroy(&pixs);
boxaDestroy(&boxar);
FREE(tab);
return pixa;
}
int main(int argc, char* argv[])
{
PIX *pixs, *pixb, *pixt;
int minb;
if (argc < 2) {
USAGE: fprintf(stderr, "Usage: %s </path/to/text-image>\n"
"\t\t[binarize-threshold] [minw:minh:maxw:maxh]\n",
strrchr(argv[0], '/') + 1);
return EINVAL;
}
if (2 < argc) { errno = 0;
minb = strtol(argv[2], NULL, 10);
if (errno < 0) {
fprintf(stderr, "strtol: %s\n", strerror(errno));
goto USAGE;
}
} else minb = 180;
if (!(pixs = pixRead(argv[1]))) ;
if (1 && (pixt = pixBackgroundNormMorph(pixs, NULL, 4, 5, 248))) {
pixDestroy(&pixs); pixs = pixt;
} else
if (0 && (pixt = pixBackgroundNorm(pixs, NULL, NULL,
10, 15, 60, 40, 248, 2, 1))) {
pixDestroy(&pixs); pixs = pixt;
}
if (1 && (pixt = pixFindSkewAndDeskew(pixs, 1, NULL, NULL))) {
pixDestroy(&pixs); pixs = pixt;
} if (0 && pixDisplay(pixs, 0, 0)) ;
if (1) { PTA *ptas, *ptad;
if (!(pixb = pixConvertTo1(pixs, minb))) ;
// pixt = pixDeskewLocal(pixs, 10, 0, 0, 0.0, 0.0, 0.0))
if (!pixGetLocalSkewTransform(pixb,
10, 0, 0, 0.0, 0.0, 0.0, &ptas, &ptad)) {
if ((pixt = pixProjectiveSampledPta(pixs,
ptad, ptas, L_BRING_IN_WHITE))) {
pixDestroy(&pixs); pixs = pixt;
} ptaDestroy(&ptas); ptaDestroy(&ptad);
} pixDestroy(&pixb);
}
if (0 && (pixt = pixGammaTRC(NULL, pixs, 1.0, 30, 180))) {
pixDestroy(&pixs); pixs = pixt;
}
if (!(pixb = pixConvertTo1(pixs, minb))) ;
if (0) { pixDestroy(&pixs); pixs = pixCopy(pixs, pixb); } // XXX:
if (1) {
BOX* box;
int i, n, j, m;
PIX *pixi, *pixl;
BOXA *boxi, *boxl;
int x, y, w, h, wid;
int X = INT_MAX, Y = INT_MAX, W = 0, H;
// XXX: do smaller(or no) pixOpenBrick
if (pixGetRegionsBinary(pixb, &pixi, &pixl, NULL, 0)) ;
boxl = pixConnComp(pixl, NULL, 4);
n = boxaGetCount(boxl);
for (i = 0; i < n; ++i) { BOXA* boxa;
box = boxaGetBox(boxl, i, L_CLONE);
boxGetGeometry(box, &x, &y, &w, &h);
if (w < 30 || h < 30 || w < h || h < (w / 40)) {
boxDestroy(&box); continue;
boxaRemoveBox(boxl, i);
}
if (x < X) X = x; if (y < Y) Y = y; if (W < w) W = w;
pixt = pixClipRectangle(pixb, box, NULL);
boxDestroy(&box);
// XXX: for English
if (0) pixt = pixDilateBrick(pixt, pixt, h >> 1, h >> 1); else
pixt = pixDilateBrick(pixt, pixt, 16 < h ? h >> 4 : 1, h << 1);
if (0 && pixDisplay(pixt, 0, 0)) ;
boxa = pixConnComp(pixt, NULL, 8);
pixDestroy(&pixt);
wid = (h * 3) >> 2;
//boxaShift(boxa, x, y);
m = boxaGetCount(boxa);
for (j = 0; j < m; ++j) {
int x0, y0, w0;
box = boxaGetBox(boxa, j, L_CLONE);
boxGetGeometry(box, &x0, &y0, &w0, NULL);
// merge adjacent 2 or 3 small boxes
if (1 && w0 < wid && (j + 1) < m) {
BOX* boxn; int xn, wn;
boxn = boxaGetBox(boxa, j + 1, L_CLONE);
boxGetGeometry(boxn, &xn, NULL, &wn, NULL);
if ((w0 = xn + wn - x0) < h) {
boxaSparseClearBox(boxa, ++j);
if (w0 < wid && (j + 1) < m) {
boxDestroy(&boxn);
boxn = boxaGetBox(boxa, j + 1, L_CLONE);
boxGetGeometry(boxn, &xn, NULL, &wn, NULL);
if ((wn = xn + wn - x0) < h) {
boxaSparseClearBox(boxa, ++j);
w0 = wn;
}
}
boxSetGeometry(box, -1, -1, w0, -1);
} boxDestroy(&boxn);
}
boxSetGeometry(box, x + x0, y + y0, -1, -1);
boxDestroy(&box);
} boxaSparseCompact(boxa);
if (1 && (pixt = pixDrawBoxa(pixs, boxa, 1, 0xff000000))) {
pixDestroy(&pixs); pixs = pixt;
} boxaDestroy(&boxa);
} H = y + h;
main(int argc,
char **argv)
{
l_int32 i, n;
l_float32 pi, angle, val;
BOX *box;
BOXA *boxa, *boxa1, *boxa2;
NUMA *na1, *na2;
PIX *pix, *pix1, *pix2, *pix3, *pixd;
PIXA *pixa1, *pixa2, *pixa3, *pixa4;
static char mainName[] = "inserttest";
#if 1
pi = 3.1415926535;
na1 = numaCreate(500);
for (i = 0; i < 500; i++) {
angle = 0.02293 * i * pi;
val = (l_float32)sin(angle);
numaAddNumber(na1, val);
}
numaWrite("/tmp/junknuma1", na1);
na2 = numaCopy(na1);
n = numaGetCount(na2);
for (i = 0; i < n; i++) {
numaGetFValue(na2, i, &val);
numaRemoveNumber(na2, i);
numaInsertNumber(na2, i, val);
}
numaWrite("/tmp/junknuma2", na2);
numaDestroy(&na1);
numaDestroy(&na2);
#endif
#if 1
pix1 = pixRead("feyn.tif");
box = boxCreate(1138, 1666, 1070, 380);
pix2 = pixClipRectangle(pix1, box, NULL);
boxDestroy(&box);
boxa1 = pixConnComp(pix2, NULL, 8);
boxaWrite("/tmp/junkboxa1", boxa1);
boxa2 = boxaCopy(boxa1, L_COPY);
n = boxaGetCount(boxa2);
for (i = 0; i < n; i++) {
box = boxaGetBox(boxa2, i, L_COPY);
boxaRemoveBox(boxa2, i);
boxaInsertBox(boxa2, i, box);
}
boxaWrite("/tmp/junkboxa2", boxa2);
pixDestroy(&pix1);
pixDestroy(&pix2);
boxaDestroy(&boxa1);
boxaDestroy(&boxa2);
#endif
#if 1
pix1 = pixRead("feyn.tif");
box = boxCreate(1138, 1666, 1070, 380);
pix2 = pixClipRectangle(pix1, box, NULL);
boxDestroy(&box);
boxa = pixConnComp(pix2, &pixa1, 8);
boxaDestroy(&boxa);
pixaWrite("/tmp/junkpixa1", pixa1);
pixa2 = pixaCopy(pixa1, L_COPY);
n = pixaGetCount(pixa2);
/* Remove and insert each one */
for (i = 0; i < n; i++) {
pix = pixaGetPix(pixa2, i, L_COPY);
box = pixaGetBox(pixa2, i, L_COPY);
pixaRemovePix(pixa2, i);
pixaInsertPix(pixa2, i, pix, box);
}
pixaWrite("/tmp/junkpixa2", pixa2);
/* Move the last to the beginning; do it n times */
pixa3 = pixaCopy(pixa2, L_COPY);
for (i = 0; i < n; i++) {
pix = pixaGetPix(pixa3, n - 1, L_CLONE);
box = pixaGetBox(pixa3, n - 1, L_CLONE);
pixaInsertPix(pixa3, 0, pix, box);
pixaRemovePix(pixa3, n);
}
pixaWrite("/tmp/junkpixa3", pixa3);
/* Move the first one to the end; do it n times */
pixa4 = pixaCopy(pixa3, L_COPY);
for (i = 0; i < n; i++) {
pix = pixaGetPix(pixa4, 0, L_CLONE);
box = pixaGetBox(pixa4, 0, L_CLONE);
pixaInsertPix(pixa4, n, pix, box); /* make sure insert works at end */
pixaRemovePix(pixa4, 0);
}
pixaWrite("/tmp/junkpixa4", pixa4);
pixDestroy(&pix1);
pixDestroy(&pix2);
pixaDestroy(&pixa1);
pixaDestroy(&pixa2);
pixaDestroy(&pixa3);
pixaDestroy(&pixa4);
#endif
return 0;
}