jboolean Java_com_googlecode_leptonica_android_Pixa_nativeWriteToFileRandomCmap(JNIEnv *env,
jclass clazz,
jint nativePixa,
jstring fileName,
jint width, jint height) {
LOGV(__FUNCTION__);
PIX *pixtemp;
PIXA *pixa = (PIXA *) nativePixa;
const char *c_fileName = env->GetStringUTFChars(fileName, NULL);
if (c_fileName == NULL) {
LOGE("could not extract fileName string!");
return JNI_FALSE;
}
if (pixaGetCount(pixa) > 0) {
pixtemp = pixaDisplayRandomCmap(pixa, (l_int32) width, (l_int32) height);
} else {
pixtemp = pixCreate((l_int32) width, (l_int32) height, 1);
}
pixWrite(c_fileName, pixtemp, IFF_BMP);
pixDestroy(&pixtemp);
env->ReleaseStringUTFChars(fileName, c_fileName);
return JNI_TRUE;
}
PIX *
MakeColorWash(l_int32 w,
l_int32 h,
l_int32 color)
{
l_int32 i, j, wpl;
l_uint32 val;
l_uint32 *data, *line;
PIX *pixd;
pixd = pixCreate(w, h, 32);
data = pixGetData(pixd);
wpl = pixGetWpl(pixd);
for (i = 0; i < h; i++) {
line = data + i * wpl;
for (j = 0; j < w; j++) {
if (color == COLOR_RED)
val = ((j * 255) / w) << L_GREEN_SHIFT |
((j * 255) / w) << L_BLUE_SHIFT |
255 << L_RED_SHIFT;
else if (color == COLOR_GREEN)
val = ((j * 255) / w) << L_RED_SHIFT |
((j * 255) / w) << L_BLUE_SHIFT |
255 << L_GREEN_SHIFT;
else
val = ((j * 255) / w) << L_RED_SHIFT |
((j * 255) / w) << L_GREEN_SHIFT |
255 << L_BLUE_SHIFT;
line[j] = val;
}
}
return pixd;
}
/*!
* pixAddRGB()
*
* Input: pixs1, pixs2 (32 bpp RGB, or colormapped)
* Return: pixd, or null on error
*
* Notes:
* (1) Clips computation to the minimum size, aligning the UL corners.
* (2) Removes any colormap to RGB, and ignores the LSB of each
* pixel word (the alpha channel).
* (3) Adds each component value, pixelwise, clipping to 255.
* (4) This is useful to combine two images where most of the
* pixels are essentially black, such as in pixPerceptualDiff().
*/
PIX *
pixAddRGB(PIX *pixs1,
PIX *pixs2)
{
l_int32 i, j, w, h, d, w2, h2, d2, wplc1, wplc2, wpld;
l_int32 rval1, gval1, bval1, rval2, gval2, bval2, rval, gval, bval;
l_uint32 *datac1, *datac2, *datad, *linec1, *linec2, *lined;
PIX *pixc1, *pixc2, *pixd;
PROCNAME("pixAddRGB");
if (!pixs1)
return (PIX *)ERROR_PTR("pixs1 not defined", procName, NULL);
if (!pixs2)
return (PIX *)ERROR_PTR("pixs2 not defined", procName, NULL);
pixGetDimensions(pixs1, &w, &h, &d);
pixGetDimensions(pixs2, &w2, &h2, &d2);
if (!pixGetColormap(pixs1) && d != 32)
return (PIX *)ERROR_PTR("pixs1 not cmapped or rgb", procName, NULL);
if (!pixGetColormap(pixs2) && d2 != 32)
return (PIX *)ERROR_PTR("pixs2 not cmapped or rgb", procName, NULL);
if (pixGetColormap(pixs1))
pixc1 = pixRemoveColormap(pixs1, REMOVE_CMAP_TO_FULL_COLOR);
else
pixc1 = pixClone(pixs1);
if (pixGetColormap(pixs2))
pixc2 = pixRemoveColormap(pixs2, REMOVE_CMAP_TO_FULL_COLOR);
else
pixc2 = pixClone(pixs2);
w = L_MIN(w, w2);
h = L_MIN(h, h2);
pixd = pixCreate(w, h, 32);
pixCopyResolution(pixd, pixs1);
datac1 = pixGetData(pixc1);
datac2 = pixGetData(pixc2);
datad = pixGetData(pixd);
wplc1 = pixGetWpl(pixc1);
wplc2 = pixGetWpl(pixc2);
wpld = pixGetWpl(pixd);
for (i = 0; i < h; i++) {
linec1 = datac1 + i * wplc1;
linec2 = datac2 + i * wplc2;
lined = datad + i * wpld;
for (j = 0; j < w; j++) {
extractRGBValues(linec1[j], &rval1, &gval1, &bval1);
extractRGBValues(linec2[j], &rval2, &gval2, &bval2);
rval = L_MIN(255, rval1 + rval2);
gval = L_MIN(255, gval1 + gval2);
bval = L_MIN(255, bval1 + bval2);
composeRGBPixel(rval, gval, bval, lined + j);
}
}
pixDestroy(&pixc1);
pixDestroy(&pixc2);
return pixd;
}
main(int argc,
char **argv)
{
l_int32 i, j, w, h, same, width, height, cx, cy;
l_uint32 val;
PIX *pixs, *pixse, *pixd1, *pixd2;
SEL *sel;
static char mainName[] = "rasterop_reg";
if (argc != 1)
return ERROR_INT(" Syntax: rasterop_reg", mainName, 1);
pixs = pixRead("feyn.tif");
for (width = 1; width <= 25; width += 3) {
for (height = 1; height <= 25; height += 4) {
cx = width / 2;
cy = height / 2;
/* Dilate using an actual sel */
sel = selCreateBrick(height, width, cy, cx, SEL_HIT);
pixd1 = pixDilate(NULL, pixs, sel);
/* Dilate using a pix as a sel */
pixse = pixCreate(width, height, 1);
pixSetAll(pixse);
pixd2 = pixCopy(NULL, pixs);
w = pixGetWidth(pixs);
h = pixGetHeight(pixs);
for (i = 0; i < h; i++) {
for (j = 0; j < w; j++) {
pixGetPixel(pixs, j, i, &val);
if (val)
pixRasterop(pixd2, j - cx, i - cy, width, height,
PIX_SRC | PIX_DST, pixse, 0, 0);
}
}
pixEqual(pixd1, pixd2, &same);
if (same == 1)
fprintf(stderr, "Correct for (%d,%d)\n", width, height);
else {
fprintf(stderr, "Error: results are different!\n");
fprintf(stderr, "SE: width = %d, height = %d\n", width, height);
pixWrite("/tmp/junkout1", pixd1, IFF_PNG);
pixWrite("/tmp/junkout2", pixd2, IFF_PNG);
return 1;
}
pixDestroy(&pixse);
pixDestroy(&pixd1);
pixDestroy(&pixd2);
selDestroy(&sel);
}
}
pixDestroy(&pixs);
return 0;
}
/*!
* pixEmbedForRotation()
*
* Input: pixs (1, 2, 4, 8, 32 bpp rgb)
* angle (radians; clockwise is positive)
* incolor (L_BRING_IN_WHITE, L_BRING_IN_BLACK)
* width (original width; use 0 to avoid embedding)
* height (original height; use 0 to avoid embedding)
* Return: pixd, or null on error
*
* Notes:
* (1) For very small rotations, just return a clone.
* (2) Generate larger image to embed pixs if necessary, and
* place the center of the input image in the center.
* (3) Rotation brings either white or black pixels in
* from outside the image. For colormapped images where
* there is no white or black, a new color is added if
* possible for these pixels; otherwise, either the
* lightest or darkest color is used. In most cases,
* the colormap will be removed prior to rotation.
* (4) The dest is to be expanded so that no image pixels
* are lost after rotation. Input of the original width
* and height allows the expansion to stop at the maximum
* required size, which is a square with side equal to
* sqrt(w*w + h*h).
* (5) For an arbitrary angle, the expansion can be found by
* considering the UL and UR corners. As the image is
* rotated, these move in an arc centered at the center of
* the image. Normalize to a unit circle by dividing by half
* the image diagonal. After a rotation of T radians, the UL
* and UR corners are at points T radians along the unit
* circle. Compute the x and y coordinates of both these
* points and take the max of absolute values; these represent
* the half width and half height of the containing rectangle.
* The arithmetic is done using formulas for sin(a+b) and cos(a+b),
* where b = T. For the UR corner, sin(a) = h/d and cos(a) = w/d.
* For the UL corner, replace a by (pi - a), and you have
* sin(pi - a) = h/d, cos(pi - a) = -w/d. The equations
* given below follow directly.
*/
PIX *
pixEmbedForRotation(PIX *pixs,
l_float32 angle,
l_int32 incolor,
l_int32 width,
l_int32 height)
{
l_int32 w, h, d, w1, h1, w2, h2, maxside, wnew, hnew, xoff, yoff, setcolor;
l_float64 sina, cosa, fw, fh;
PIX *pixd;
PROCNAME("pixEmbedForRotation");
if (!pixs)
return (PIX *)ERROR_PTR("pixs not defined", procName, NULL);
if (incolor != L_BRING_IN_WHITE && incolor != L_BRING_IN_BLACK)
return (PIX *)ERROR_PTR("invalid incolor", procName, NULL);
if (L_ABS(angle) < MIN_ANGLE_TO_ROTATE)
return pixClone(pixs);
/* Test if big enough to hold any rotation of the original image */
pixGetDimensions(pixs, &w, &h, &d);
maxside = (l_int32)(sqrt((l_float64)(width * width) +
(l_float64)(height * height)) + 0.5);
if (w >= maxside && h >= maxside) /* big enough */
return pixClone(pixs);
/* Find the new sizes required to hold the image after rotation.
* Note that the new dimensions must be at least as large as those
* of pixs, because we're rasterop-ing into it before rotation. */
cosa = cos(angle);
sina = sin(angle);
fw = (l_float64)w;
fh = (l_float64)h;
w1 = (l_int32)(L_ABS(fw * cosa - fh * sina) + 0.5);
w2 = (l_int32)(L_ABS(-fw * cosa - fh * sina) + 0.5);
h1 = (l_int32)(L_ABS(fw * sina + fh * cosa) + 0.5);
h2 = (l_int32)(L_ABS(-fw * sina + fh * cosa) + 0.5);
wnew = L_MAX(w, L_MAX(w1, w2));
hnew = L_MAX(h, L_MAX(h1, h2));
if ((pixd = pixCreate(wnew, hnew, d)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
pixCopyResolution(pixd, pixs);
pixCopyColormap(pixd, pixs);
pixCopySpp(pixd, pixs);
pixCopyText(pixd, pixs);
xoff = (wnew - w) / 2;
yoff = (hnew - h) / 2;
/* Set background to color to be rotated in */
setcolor = (incolor == L_BRING_IN_BLACK) ? L_SET_BLACK : L_SET_WHITE;
pixSetBlackOrWhite(pixd, setcolor);
/* Rasterop automatically handles all 4 channels for rgba */
pixRasterop(pixd, xoff, yoff, w, h, PIX_SRC, pixs, 0, 0);
return pixd;
}
jint Java_com_googlecode_leptonica_android_ReadFile_nativeReadBitmap(JNIEnv *env,
jclass clazz,
jobject bitmap) {
//LOGV(__FUNCTION__);
l_int32 w, h, d;
AndroidBitmapInfo info;
void* pixels;
int ret;
if ((ret = AndroidBitmap_getInfo(env, bitmap, &info)) < 0) {
LOGE("AndroidBitmap_getInfo() failed ! error=%d", ret);
return JNI_FALSE;
}
if (info.format != ANDROID_BITMAP_FORMAT_RGBA_8888) {
LOGE("Bitmap format is not RGBA_8888 !");
return JNI_FALSE;
}
if ((ret = AndroidBitmap_lockPixels(env, bitmap, &pixels)) < 0) {
LOGE("AndroidBitmap_lockPixels() failed ! error=%d", ret);
return JNI_FALSE;
}
PIX *pixd = pixCreate(info.width, info.height, 8);
l_uint32 *src = (l_uint32 *) pixels;
l_int32 srcWpl = (info.stride / 4);
l_uint32 *dst = pixGetData(pixd);
l_int32 dstWpl = pixGetWpl(pixd);
l_uint8 a, r, g, b, pixel8;
for (int y = 0; y < info.height; y++) {
l_uint32 *dst_line = dst + (y * dstWpl);
l_uint32 *src_line = src + (y * srcWpl);
for (int x = 0; x < info.width; x++) {
// Get pixel from RGBA_8888
r = *src_line >> SK_R32_SHIFT;
g = *src_line >> SK_G32_SHIFT;
b = *src_line >> SK_B32_SHIFT;
a = *src_line >> SK_A32_SHIFT;
pixel8 = (l_uint8) ((r + g + b) / 3);
// Set pixel to LUMA_8
SET_DATA_BYTE(dst_line, x, pixel8);
// Move to the next pixel
src_line++;
}
}
AndroidBitmap_unlockPixels(env, bitmap);
return (jint) pixd;
}
/*!
* pixRasteropHip()
*
* Input: pixd (in-place operation)
* by (top of horizontal band)
* bh (height of horizontal band)
* hshift (horizontal shift of band; hshift > 0 is to right)
* incolor (L_BRING_IN_WHITE, L_BRING_IN_BLACK)
* Return: 0 if OK; 1 on error
*
* Notes:
* (1) This rasterop translates a horizontal band of the
* image either left or right, bringing in either white
* or black pixels from outside the image.
* (2) The horizontal band extends the full width of pixd.
* (3) If a colormap exists, the nearest color to white or black
* is brought in.
*/
l_int32
pixRasteropHip(PIX *pixd,
l_int32 by,
l_int32 bh,
l_int32 hshift,
l_int32 incolor)
{
l_int32 w, h, d, index, op;
PIX *pixt;
PIXCMAP *cmap;
PROCNAME("pixRasteropHip");
if (!pixd)
return ERROR_INT("pixd not defined", procName, 1);
if (incolor != L_BRING_IN_WHITE && incolor != L_BRING_IN_BLACK)
return ERROR_INT("invalid value for incolor", procName, 1);
if (bh <= 0)
return ERROR_INT("bh must be > 0", procName, 1);
if (hshift == 0)
return 0;
pixGetDimensions(pixd, &w, &h, &d);
rasteropHipLow(pixGetData(pixd), h, d, pixGetWpl(pixd), by, bh, hshift);
cmap = pixGetColormap(pixd);
if (!cmap) {
if ((d == 1 && incolor == L_BRING_IN_BLACK) ||
(d > 1 && incolor == L_BRING_IN_WHITE))
op = PIX_SET;
else
op = PIX_CLR;
/* Set the pixels brought in at left or right */
if (hshift > 0)
pixRasterop(pixd, 0, by, hshift, bh, op, NULL, 0, 0);
else /* hshift < 0 */
pixRasterop(pixd, w + hshift, by, -hshift, bh, op, NULL, 0, 0);
return 0;
}
/* Get the nearest index and fill with that */
if (incolor == L_BRING_IN_BLACK)
pixcmapGetRankIntensity(cmap, 0.0, &index);
else /* white */
pixcmapGetRankIntensity(cmap, 1.0, &index);
pixt = pixCreate(L_ABS(hshift), bh, d);
pixSetAllArbitrary(pixt, index);
if (hshift > 0)
pixRasterop(pixd, 0, by, hshift, bh, PIX_SRC, pixt, 0, 0);
else /* hshift < 0 */
pixRasterop(pixd, w + hshift, by, -hshift, bh, PIX_SRC, pixt, 0, 0);
pixDestroy(&pixt);
return 0;
}
jint Java_com_googlecode_leptonica_android_ReadFile_nativeReadBitmap(JNIEnv *env,
jclass clazz,
jobject bitmap) {
LOGV(__FUNCTION__);
l_int32 w, h, d;
AndroidBitmapInfo info;
void* pixels;
int ret;
if ((ret = AndroidBitmap_getInfo(env, bitmap, &info)) < 0) {
LOGE("AndroidBitmap_getInfo() failed ! error=%d", ret);
return JNI_FALSE;
}
if (info.format != ANDROID_BITMAP_FORMAT_RGBA_8888) {
LOGE("Bitmap format is not RGBA_8888 !");
return JNI_FALSE;
}
if ((ret = AndroidBitmap_lockPixels(env, bitmap, &pixels)) < 0) {
LOGE("AndroidBitmap_lockPixels() failed ! error=%d", ret);
return JNI_FALSE;
}
PIX *pix = pixCreate(info.width, info.height, 8);
l_uint8 *src = (l_uint8 *) pixels;
l_uint8 *dst = (l_uint8 *) pixGetData(pixs);
l_int32 srcBpl = info.stride;
l_int32 dstBpl = 4 * pixGetWpl(pixs);
l_uint8 a, r, g, b;
SET_PIXEL = SET_PIXEL_8;
for (int y = 0; y < info.height; y++) {
l_uint8 *dst_line = dst + (y * dstBpl);
l_uint8 *src_line = src + (y * srcBpl);
for (int x = 0; x < info.width; x++) {
// Get pixel from ARGB_8888
r = *src_line++;
g = *src_line++;
b = *src_line++;
a = *src_line++;
// Set pixel to LUMA_8
*dst_line = (l_uint8) ((r + g + b) / 3);
}
}
AndroidBitmap_unlockPixels(env, bitmap);
return (jint) pix;
}
/*!
* \brief pixColorSegmentCluster()
*
* \param[in] pixs 32 bpp; 24-bit color
* \param[in] maxdist max euclidean dist to existing cluster
* \param[in] maxcolors max number of colors allowed in first pass
* \param[in] debugflag 1 for debug output; 0 otherwise
* \return pixd 8 bit with colormap, or NULL on error
*
* <pre>
* Notes:
* (1) This is phase 1. See description in pixColorSegment().
* (2) Greedy unsupervised classification. If the limit 'maxcolors'
* is exceeded, the computation is repeated with a larger
* allowed cluster size.
* (3) On each successive iteration, 'maxdist' is increased by a
* constant factor. See comments in pixColorSegment() for
* a guideline on parameter selection.
* Note that the diagonal of the 8-bit rgb color cube is about
* 440, so for 'maxdist' = 440, you are guaranteed to get 1 color!
* </pre>
*/
PIX *
pixColorSegmentCluster(PIX *pixs,
l_int32 maxdist,
l_int32 maxcolors,
l_int32 debugflag)
{
l_int32 w, h, newmaxdist, ret, niters, ncolors, success;
PIX *pixd;
PIXCMAP *cmap;
PROCNAME("pixColorSegmentCluster");
if (!pixs)
return (PIX *)ERROR_PTR("pixs not defined", procName, NULL);
if (pixGetDepth(pixs) != 32)
return (PIX *)ERROR_PTR("must be rgb color", procName, NULL);
pixGetDimensions(pixs, &w, &h, NULL);
if ((pixd = pixCreate(w, h, 8)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
cmap = pixcmapCreate(8);
pixSetColormap(pixd, cmap);
pixCopyResolution(pixd, pixs);
newmaxdist = maxdist;
niters = 0;
success = TRUE;
while (1) {
ret = pixColorSegmentTryCluster(pixd, pixs, newmaxdist,
maxcolors, debugflag);
niters++;
if (!ret) {
ncolors = pixcmapGetCount(cmap);
if (debugflag)
L_INFO("Success with %d colors after %d iters\n", procName,
ncolors, niters);
break;
}
if (niters == MAX_ALLOWED_ITERATIONS) {
L_WARNING("too many iters; newmaxdist = %d\n",
procName, newmaxdist);
success = FALSE;
break;
}
newmaxdist = (l_int32)(DIST_EXPAND_FACT * (l_float32)newmaxdist);
}
if (!success) {
pixDestroy(&pixd);
return (PIX *)ERROR_PTR("failure in phase 1", procName, NULL);
}
return pixd;
}
/*!
* pixaDisplayRandomCmap()
*
* Input: pixa (of 1 bpp components, with boxa)
* w, h (if set to 0, determines the size from the
* b.b. of the components in pixa)
* Return: pix (8 bpp, cmapped, with random colors on the components),
* or null on error
*
* Notes:
* (1) This uses the boxes to place each pix in the rendered composite.
* (2) By default, the background color is: black, cmap index 0.
* This can be changed by pixcmapResetColor()
*/
PIX *
pixaDisplayRandomCmap(PIXA *pixa,
l_int32 w,
l_int32 h)
{
l_int32 i, n, d, index, xb, yb, wb, hb;
BOXA *boxa;
PIX *pixs, *pixt, *pixd;
PIXCMAP *cmap;
PROCNAME("pixaDisplayRandomCmap");
if (!pixa)
return (PIX *)ERROR_PTR("pixa not defined", procName, NULL);
n = pixaGetCount(pixa);
if (n == 0)
return (PIX *)ERROR_PTR("no components", procName, NULL);
/* Use the first pix in pixa to verify depth is 1 bpp */
pixs = pixaGetPix(pixa, 0, L_CLONE);
d = pixGetDepth(pixs);
pixDestroy(&pixs);
if (d != 1)
return (PIX *)ERROR_PTR("components not 1 bpp", procName, NULL);
/* If w and h not input, determine the minimum size required
* to contain the origin and all c.c. */
if (w == 0 || h == 0) {
boxa = pixaGetBoxa(pixa, L_CLONE);
boxaGetExtent(boxa, &w, &h, NULL);
boxaDestroy(&boxa);
}
/* Set up an 8 bpp dest pix, with a colormap with 254 random colors */
if ((pixd = pixCreate(w, h, 8)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
cmap = pixcmapCreateRandom(8, 1, 1);
pixSetColormap(pixd, cmap);
/* Color each component and blit it in */
for (i = 0; i < n; i++) {
index = 1 + (i % 254);
pixaGetBoxGeometry(pixa, i, &xb, &yb, &wb, &hb);
pixs = pixaGetPix(pixa, i, L_CLONE);
pixt = pixConvert1To8(NULL, pixs, 0, index);
pixRasterop(pixd, xb, yb, wb, hb, PIX_PAINT, pixt, 0, 0);
pixDestroy(&pixs);
pixDestroy(&pixt);
}
return pixd;
}
/*!
* pixMaxDynamicRange()
*
* Input: pixs (4, 8, 16 or 32 bpp source)
* type (L_LINEAR_SCALE or L_LOG_SCALE)
* Return: pixd (8 bpp), or null on error
*
* Notes:
* (1) Scales pixel values to fit maximally within the dest 8 bpp pixd
* (2) Uses a LUT for log scaling
*/
PIX *
pixMaxDynamicRange(PIX *pixs,
l_int32 type)
{
l_uint8 dval;
l_int32 i, j, w, h, d, wpls, wpld, max, sval;
l_uint32 *datas, *datad;
l_uint32 word;
l_uint32 *lines, *lined;
l_float32 factor;
l_float32 *tab;
PIX *pixd;
PROCNAME("pixMaxDynamicRange");
if (!pixs)
return (PIX *)ERROR_PTR("pixs not defined", procName, NULL);
d = pixGetDepth(pixs);
if (d != 4 && d != 8 && d != 16 && d != 32)
return (PIX *)ERROR_PTR("pixs not in {4,8,16,32} bpp", procName, NULL);
if (type != L_LINEAR_SCALE && type != L_LOG_SCALE)
return (PIX *)ERROR_PTR("invalid type", procName, NULL);
pixGetDimensions(pixs, &w, &h, NULL);
if ((pixd = pixCreate(w, h, 8)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
pixCopyResolution(pixd, pixs);
datas = pixGetData(pixs);
datad = pixGetData(pixd);
wpls = pixGetWpl(pixs);
wpld = pixGetWpl(pixd);
/* Get max */
max = 0;
for (i = 0; i < h; i++) {
lines = datas + i * wpls;
for (j = 0; j < wpls; j++) {
word = *(lines + j);
if (d == 4) {
max = L_MAX(max, word >> 28);
max = L_MAX(max, (word >> 24) & 0xf);
max = L_MAX(max, (word >> 20) & 0xf);
max = L_MAX(max, (word >> 16) & 0xf);
max = L_MAX(max, (word >> 12) & 0xf);
max = L_MAX(max, (word >> 8) & 0xf);
max = L_MAX(max, (word >> 4) & 0xf);
max = L_MAX(max, word & 0xf);
} else if (d == 8) {
max = L_MAX(max, word >> 24);
max = L_MAX(max, (word >> 16) & 0xff);
max = L_MAX(max, (word >> 8) & 0xff);
max = L_MAX(max, word & 0xff);
} else if (d == 16) {
/*!
* \brief pixExpandBinaryReplicate()
*
* \param[in] pixs 1 bpp
* \param[in] xfact integer scale factor for horiz. replicative expansion
* \param[in] yfact integer scale factor for vertical replicative expansion
* \return pixd scaled up, or NULL on error
*/
PIX *
pixExpandBinaryReplicate(PIX *pixs,
l_int32 xfact,
l_int32 yfact)
{
l_int32 w, h, d, wd, hd, wpls, wpld, i, j, k, start;
l_uint32 *datas, *datad, *lines, *lined;
PIX *pixd;
PROCNAME("pixExpandBinaryReplicate");
if (!pixs)
return (PIX *)ERROR_PTR("pixs not defined", procName, NULL);
pixGetDimensions(pixs, &w, &h, &d);
if (d != 1)
return (PIX *)ERROR_PTR("pixs not binary", procName, NULL);
if (xfact <= 0 || yfact <= 0)
return (PIX *)ERROR_PTR("invalid scale factor: <= 0", procName, NULL);
if (xfact == yfact) {
if (xfact == 1)
return pixCopy(NULL, pixs);
if (xfact == 2 || xfact == 4 || xfact == 8 || xfact == 16)
return pixExpandBinaryPower2(pixs, xfact);
}
wpls = pixGetWpl(pixs);
datas = pixGetData(pixs);
wd = xfact * w;
hd = yfact * h;
if ((pixd = pixCreate(wd, hd, 1)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
pixCopyResolution(pixd, pixs);
pixScaleResolution(pixd, (l_float32)xfact, (l_float32)yfact);
wpld = pixGetWpl(pixd);
datad = pixGetData(pixd);
for (i = 0; i < h; i++) {
lines = datas + i * wpls;
lined = datad + yfact * i * wpld;
for (j = 0; j < w; j++) { /* replicate pixels on a single line */
if (GET_DATA_BIT(lines, j)) {
start = xfact * j;
for (k = 0; k < xfact; k++)
SET_DATA_BIT(lined, start + k);
}
}
for (k = 1; k < yfact; k++) /* replicate the line */
memcpy(lined + k * wpld, lined, 4 * wpld);
}
return pixd;
}
char* ocr_bitmap(void* arg, png_color *palette,png_byte *alpha, unsigned char* indata,int w, int h)
{
PIX *pix = NULL;
PIX *cpix = NULL;
char*text_out= NULL;
int i,j,index;
unsigned int wpl;
unsigned int *data,*ppixel;
BOOL tess_ret = FALSE;
struct ocrCtx* ctx = arg;
pix = pixCreate(w, h, 32);
if(pix == NULL)
{
return NULL;
}
wpl = pixGetWpl(pix);
data = pixGetData(pix);
#if LEPTONICA_VERSION > 69
pixSetSpp(pix, 4);
#endif
for (i = 0; i < h; i++)
{
ppixel = data + i * wpl;
for (j = 0; j < w; j++)
{
index = indata[i * w + (j)];
composeRGBPixel(palette[index].red, palette[index].green,palette[index].blue, ppixel);
SET_DATA_BYTE(ppixel, L_ALPHA_CHANNEL,alpha[index]);
ppixel++;
}
}
ignore_alpha_at_edge(alpha, indata, w, h, pix, &cpix);
#ifdef OCR_DEBUG
{
char str[128] = "";
static int i = 0;
sprintf(str,"temp/file_c_%d.png",i);
pixWrite(str, cpix, IFF_PNG);
i++;
}
#endif
TessBaseAPISetImage2(ctx->api, cpix);
tess_ret = TessBaseAPIRecognize(ctx->api, NULL);
if( tess_ret != 0)
printf("\nsomething messy\n");
text_out = TessBaseAPIGetUTF8Text(ctx->api);
pixDestroy(&pix);
pixDestroy(&cpix);
return text_out;
}
// Make a Pix of the correct scaled size for the TraceOutline functions.
Pix* GridReducedPix(const TBOX& box, int gridsize,
ICOORD bleft, int* left, int* bottom) {
// Compute grid bounds of the outline and pad all round by 1.
int grid_left = (box.left() - bleft.x()) / gridsize - 1;
int grid_bottom = (box.bottom() - bleft.y()) / gridsize - 1;
int grid_right = (box.right() - bleft.x()) / gridsize + 1;
int grid_top = (box.top() - bleft.y()) / gridsize + 1;
*left = grid_left;
*bottom = grid_bottom;
return pixCreate(grid_right - grid_left + 1,
grid_top - grid_bottom + 1,
1);
}
/*!
* wshedCreate()
*
* Input: pixs (8 bpp source)
* pixm (1 bpp 'marker' seed)
* mindepth (minimum depth; anything less is not saved)
* debugflag (1 for debug output)
* Return: WShed, or null on error
*
* Notes:
* (1) It is not necessary for the fg pixels in the seed image
* be at minima, or that they be isolated. We extract a
* single pixel from each connected component, and a seed
* anywhere in a watershed will eventually label the watershed
* when the filling level reaches it.
* (2) Set mindepth to some value to ignore noise in pixs that
* can create small local minima. Any watershed shallower
* than mindepth, even if it has a seed, will not be saved;
* It will either be incorporated in another watershed or
* eliminated.
*/
L_WSHED *
wshedCreate(PIX *pixs,
PIX *pixm,
l_int32 mindepth,
l_int32 debugflag)
{
l_int32 w, h;
L_WSHED *wshed;
PROCNAME("wshedCreate");
if (!pixs)
return (L_WSHED *)ERROR_PTR("pixs is not defined", procName, NULL);
if (pixGetDepth(pixs) != 8)
return (L_WSHED *)ERROR_PTR("pixs is not 8 bpp", procName, NULL);
if (!pixm)
return (L_WSHED *)ERROR_PTR("pixm is not defined", procName, NULL);
if (pixGetDepth(pixm) != 1)
return (L_WSHED *)ERROR_PTR("pixm is not 1 bpp", procName, NULL);
pixGetDimensions(pixs, &w, &h, NULL);
if (pixGetWidth(pixm) != w || pixGetHeight(pixm) != h)
return (L_WSHED *)ERROR_PTR("pixs/m sizes are unequal", procName, NULL);
if ((wshed = (L_WSHED *)CALLOC(1, sizeof(L_WSHED))) == NULL)
return (L_WSHED *)ERROR_PTR("wshed not made", procName, NULL);
wshed->pixs = pixClone(pixs);
wshed->pixm = pixClone(pixm);
wshed->mindepth = L_MAX(1, mindepth);
wshed->pixlab = pixCreate(w, h, 32);
pixSetAllArbitrary(wshed->pixlab, MAX_LABEL_VALUE);
wshed->pixt = pixCreate(w, h, 1);
wshed->lines8 = pixGetLinePtrs(pixs, NULL);
wshed->linem1 = pixGetLinePtrs(pixm, NULL);
wshed->linelab32 = pixGetLinePtrs(wshed->pixlab, NULL);
wshed->linet1 = pixGetLinePtrs(wshed->pixt, NULL);
wshed->debug = debugflag;
return wshed;
}
/*!
* pixConnCompTransform()
*
* Input: pixs (1 bpp)
* connect (connectivity: 4 or 8)
* depth (of pixd: 8 or 16 bpp; use 0 for auto determination)
* Return: pixd (8 or 16 bpp), or null on error
*
* Notes:
* (1) pixd is 8 or 16 bpp, and the pixel values label the fg component,
* starting with 1. Pixels in the bg are labelled 0.
* (2) If @depth = 0, the depth of pixd is 8 if the number of c.c.
* is less than 254, and 16 otherwise.
* (3) If @depth = 8, the assigned label for the n-th component is
* 1 + n % 254. We use mod 254 because 0 is uniquely assigned
* to black: e.g., see pixcmapCreateRandom(). Likewise,
* if @depth = 16, the assigned label uses mod(2^16 - 2).
*/
PIX *
pixConnCompTransform(PIX *pixs,
l_int32 connect,
l_int32 depth)
{
l_int32 i, n, index, w, h, xb, yb, wb, hb;
BOXA *boxa;
PIX *pix1, *pix2, *pixd;
PIXA *pixa;
PROCNAME("pixConnCompTransform");
if (!pixs || pixGetDepth(pixs) != 1)
return (PIX *)ERROR_PTR("pixs undefined or not 1 bpp", procName, NULL);
if (connect != 4 && connect != 8)
return (PIX *)ERROR_PTR("connectivity must be 4 or 8", procName, NULL);
if (depth != 0 && depth != 8 && depth != 16)
return (PIX *)ERROR_PTR("depth must be 0, 8 or 16", procName, NULL);
boxa = pixConnComp(pixs, &pixa, connect);
n = pixaGetCount(pixa);
boxaDestroy(&boxa);
pixGetDimensions(pixs, &w, &h, NULL);
if (depth == 0) {
depth = (n < 254) ? 8 : 16;
}
pixd = pixCreate(w, h, depth);
if (n == 0) { /* no fg */
pixaDestroy(&pixa);
return pixd;
}
/* Label each component and blit it in */
for (i = 0; i < n; i++) {
pixaGetBoxGeometry(pixa, i, &xb, &yb, &wb, &hb);
pix1 = pixaGetPix(pixa, i, L_CLONE);
if (depth == 8) {
index = 1 + (i % 254);
pix2 = pixConvert1To8(NULL, pix1, 0, index);
} else { /* depth == 16 */
index = 1 + (i % 0xfffe);
pix2 = pixConvert1To16(NULL, pix1, 0, index);
}
pixRasterop(pixd, xb, yb, wb, hb, PIX_PAINT, pix2, 0, 0);
pixDestroy(&pix1);
pixDestroy(&pix2);
}
pixaDestroy(&pixa);
return pixd;
}
static PIX *
DisplayBoxa(BOXA *boxa)
{
l_int32 w, h;
BOX *box;
PIX *pix1, *pix2, *pix3;
PIXA *pixa;
pixa = pixaCreate(2);
boxaGetExtent(boxa, &w, &h, &box);
pix1 = pixCreate(w, h, 1);
pixMaskBoxa(pix1, pix1, boxa, L_SET_PIXELS);
pixaAddPix(pixa, pix1, L_INSERT);
pix2 = pixCreate(w, h, 32);
pixSetAll(pix2);
pixRenderBoxaArb(pix2, boxa, 2, 0, 255, 0);
pixRenderBoxArb(pix2, box, 3, 255, 0, 0);
pixaAddPix(pixa, pix2, L_INSERT);
pix3 = pixaDisplayTiledInRows(pixa, 32, 1000, 1.0, 0, 30, 2);
boxDestroy(&box);
pixaDestroy(&pixa);
return pix3;
}
/*!
* boxaGetCoverage()
*
* Input: boxa
* wc, hc (dimensions of overall clipping rectangle with UL
* corner at (0, 0) that is covered by the boxes.
* exactflag (1 for guaranteeing an exact result; 0 for getting
* an exact result only if the boxes do not overlap)
* &fract (<return> sum of box area as fraction of w * h)
* Return: 0 if OK, 1 on error
*
* Notes:
* (1) The boxes in boxa are clipped to the input rectangle.
* (2) * When @exactflag == 1, we generate a 1 bpp pix of size
* wc x hc, paint all the boxes black, and count the fg pixels.
* This can take 1 msec on a large page with many boxes.
* * When @exactflag == 0, we clip each box to the wc x hc region
* and sum the resulting areas. This is faster.
* * The results are the same when none of the boxes overlap
* within the wc x hc region.
*/
l_int32
boxaGetCoverage(BOXA *boxa,
l_int32 wc,
l_int32 hc,
l_int32 exactflag,
l_float32 *pfract)
{
l_int32 i, n, x, y, w, h, sum;
BOX *box, *boxc;
PIX *pixt;
PROCNAME("boxaGetCoverage");
if (!pfract)
return ERROR_INT("&fract not defined", procName, 1);
*pfract = 0.0;
if (!boxa)
return ERROR_INT("boxa not defined", procName, 1);
n = boxaGetCount(boxa);
if (n == 0)
return ERROR_INT("no boxes in boxa", procName, 1);
if (exactflag == 0) { /* quick and dirty */
sum = 0;
for (i = 0; i < n; i++) {
box = boxaGetBox(boxa, i, L_CLONE);
if ((boxc = boxClipToRectangle(box, wc, hc)) != NULL) {
boxGetGeometry(boxc, NULL, NULL, &w, &h);
sum += w * h;
boxDestroy(&boxc);
}
boxDestroy(&box);
}
}
else { /* slower and exact */
pixt = pixCreate(wc, hc, 1);
for (i = 0; i < n; i++) {
box = boxaGetBox(boxa, i, L_CLONE);
boxGetGeometry(box, &x, &y, &w, &h);
pixRasterop(pixt, x, y, w, h, PIX_SET, NULL, 0, 0);
boxDestroy(&box);
}
pixCountPixels(pixt, &sum, NULL);
pixDestroy(&pixt);
}
*pfract = (l_float32)sum / (l_float32)(wc * hc);
return 0;
}
/*!
* pixExpandBinaryReplicate()
*
* Input: pixs (1 bpp)
* factor (integer scale factor for replicative expansion)
* Return: pixd (scaled up), or null on error
*/
PIX *
pixExpandBinaryReplicate(PIX *pixs,
l_int32 factor)
{
l_int32 w, h, d, wd, hd, wpls, wpld, i, j, k, start;
l_uint32 *datas, *datad, *lines, *lined;
PIX *pixd;
PROCNAME("pixExpandBinaryReplicate");
if (!pixs)
return (PIX *)ERROR_PTR("pixs not defined", procName, NULL);
pixGetDimensions(pixs, &w, &h, &d);
if (d != 1)
return (PIX *)ERROR_PTR("pixs not binary", procName, NULL);
if (factor <= 0)
return (PIX *)ERROR_PTR("factor <= 0; invalid", procName, NULL);
if (factor == 1)
return pixCopy(NULL, pixs);
if (factor == 2 || factor == 4 || factor == 8 || factor == 16)
return pixExpandBinaryPower2(pixs, factor);
wpls = pixGetWpl(pixs);
datas = pixGetData(pixs);
wd = factor * w;
hd = factor * h;
if ((pixd = pixCreate(wd, hd, 1)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
pixCopyResolution(pixd, pixs);
pixScaleResolution(pixd, (l_float32)factor, (l_float32)factor);
wpld = pixGetWpl(pixd);
datad = pixGetData(pixd);
for (i = 0; i < h; i++) {
lines = datas + i * wpls;
lined = datad + factor * i * wpld;
for (j = 0; j < w; j++) {
if (GET_DATA_BIT(lines, j)) {
start = factor * j;
for (k = 0; k < factor; k++)
SET_DATA_BIT(lined, start + k);
}
}
for (k = 1; k < factor; k++)
memcpy(lined + k * wpld, lined, 4 * wpld);
}
return pixd;
}
/*!
* fpixDisplayMaxDynamicRange()
*
* Input: fpixs
* Return: pixd (8 bpp), or null on error
*/
PIX *
fpixDisplayMaxDynamicRange(FPIX *fpixs)
{
l_uint8 dval;
l_int32 i, j, w, h, wpls, wpld;
l_float32 factor, sval, maxval;
l_float32 *lines, *datas;
l_uint32 *lined, *datad;
PIX *pixd;
PROCNAME("fpixDisplayMaxDynamicRange");
if (!fpixs)
return (PIX *)ERROR_PTR("fpixs not defined", procName, NULL);
fpixGetDimensions(fpixs, &w, &h);
datas = fpixGetData(fpixs);
wpls = fpixGetWpl(fpixs);
maxval = 0.0;
for (i = 0; i < h; i++) {
lines = datas + i * wpls;
for (j = 0; j < w; j++) {
sval = *(lines + j);
if (sval > maxval)
maxval = sval;
}
}
pixd = pixCreate(w, h, 8);
if (maxval == 0.0)
return pixd; /* all pixels are 0 */
datad = pixGetData(pixd);
wpld = pixGetWpl(pixd);
factor = 255. / maxval;
for (i = 0; i < h; i++) {
lines = datas + i * wpls;
lined = datad + i * wpld;
for (j = 0; j < w; j++) {
sval = *(lines + j);
if (sval < 0.0) sval = 0.0;
dval = (l_uint8)(factor * sval + 0.5);
SET_DATA_BYTE(lined, j, dval);
}
}
return pixd;
}
/*!
* pixaDisplayUnsplit()
*
* Input: pixa
* nx (number of mosaic cells horizontally)
* ny (number of mosaic cells vertically)
* borderwidth (of added border on all sides)
* bordercolor (in our RGBA format: 0xrrggbbaa)
* Return: pix of tiled images, or null on error
*
* Notes:
* (1) This is a logical inverse of pixaSplitPix(). It
* constructs a pix from a mosaic of tiles, all of equal size.
* (2) For added generality, a border of arbitrary color can
* be added to each of the tiles.
* (3) In use, pixa will typically have either been generated
* from pixaSplitPix() or will derived from a pixa that
* was so generated.
* (4) All pix in the pixa must be of equal depth, and, if
* colormapped, have the same colormap.
*/
PIX *
pixaDisplayUnsplit(PIXA *pixa,
l_int32 nx,
l_int32 ny,
l_int32 borderwidth,
l_uint32 bordercolor)
{
l_int32 w, h, d, wt, ht;
l_int32 i, j, k, x, y, n;
PIX *pixt, *pixd;
PROCNAME("pixaDisplayUnsplit");
if (!pixa)
return (PIX *)ERROR_PTR("pixa not defined", procName, NULL);
if (nx <= 0 || ny <= 0)
return (PIX *)ERROR_PTR("nx and ny must be > 0", procName, NULL);
if ((n = pixaGetCount(pixa)) == 0)
return (PIX *)ERROR_PTR("no components", procName, NULL);
if (n != nx * ny)
return (PIX *)ERROR_PTR("n != nx * ny", procName, NULL);
borderwidth = L_MAX(0, borderwidth);
pixaGetPixDimensions(pixa, 0, &wt, &ht, &d);
w = nx * (wt + 2 * borderwidth);
h = ny * (ht + 2 * borderwidth);
if ((pixd = pixCreate(w, h, d)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
pixt = pixaGetPix(pixa, 0, L_CLONE);
pixCopyColormap(pixd, pixt);
pixDestroy(&pixt);
if (borderwidth > 0)
pixSetAllArbitrary(pixd, bordercolor);
y = borderwidth;
for (i = 0, k = 0; i < ny; i++) {
x = borderwidth;
for (j = 0; j < nx; j++, k++) {
pixt = pixaGetPix(pixa, k, L_CLONE);
pixRasterop(pixd, x, y, wt, ht, PIX_SRC, pixt, 0, 0);
pixDestroy(&pixt);
x += wt + 2 * borderwidth;
}
y += ht + 2 * borderwidth;
}
return pixd;
}
/*!
* pixDeserializeFromMemory()
*
* Input: data (serialized data in memory)
* nbytes (number of bytes in data string)
* Return: pix, or NULL on error
*
* Notes:
* (1) See pixSerializeToMemory() for the binary format.
*/
PIX *
pixDeserializeFromMemory(const l_uint32 *data,
size_t nbytes)
{
char *id;
l_int32 w, h, d, imdatasize, ncolors;
l_uint32 *imdata; /* data in pix raster */
PIX *pixd;
PIXCMAP *cmap;
PROCNAME("pixDeserializeFromMemory");
if (!data)
return (PIX *)ERROR_PTR("data not defined", procName, NULL);
if (nbytes < 28)
return (PIX *)ERROR_PTR("invalid data", procName, NULL);
id = (char *)data;
if (id[0] != 's' || id[1] != 'p' || id[2] != 'i' || id[3] != 'x')
return (PIX *)ERROR_PTR("invalid id string", procName, NULL);
w = data[1];
h = data[2];
d = data[3];
if ((pixd = pixCreate(w, h, d)) == NULL)
return (PIX *)ERROR_PTR("pix not made", procName, NULL);
ncolors = data[5];
if (ncolors > 0) {
cmap = pixcmapDeserializeFromMemory((l_uint8 *)(&data[6]), 4, ncolors);
if (!cmap)
return (PIX *)ERROR_PTR("cmap not made", procName, NULL);
pixSetColormap(pixd, cmap);
}
imdata = pixGetData(pixd);
imdatasize = nbytes - 24 - 4 * ncolors - 4;
if (imdatasize != data[6 + ncolors])
L_ERROR("imdatasize is inconsistent with nbytes\n", procName);
memcpy((char *)imdata, (char *)(data + 7 + ncolors), imdatasize);
#if DEBUG_SERIALIZE
fprintf(stderr, "Deserialize: "
"raster size = %d, ncolors in cmap = %d, total bytes = %lu\n",
imdatasize, ncolors, nbytes);
#endif /* DEBUG_SERIALIZE */
return pixd;
}
int main(int argc, char** argv) {
if(argc != 5 && argc != 4) syntax();
int startarg = 1;
int x=INT_MAX,y=INT_MAX,xmax=INT_MIN,ymax=INT_MIN,w=-1,h=-1;
if(strncmp(argv[1], "-dim=", 5) == 0) {
sscanf(argv[1]+5, "%d,%d,%d,%d", &x, &y, &w, &h);
startarg++;
}
const char* pic1 = argv[startarg++];
const char* pic2 = argv[startarg++];
const char* out = argv[startarg++];
struct Pix* p1 = pixRead(pic1);
struct Pix* p2 = pixRead(pic2);
if(p1->h != p2->h || p1->w != p2->w) {
dprintf(2, "error: both pics need same dimensions!");
exit(1);
}
struct Pix* p1b = pixConvertTo32(p1);
struct Pix* p2b = pixConvertTo32(p2);
int X, Y, W = p1->w, H = p1->h;
if(x == INT_MAX) {
for(Y = 0; Y < H; Y++) {
for(X = 0; X < W; X++) {
if(((uint32_t*)p1b->data)[Y * W + X] != ((uint32_t*)p2b->data)[Y * W + X]) {
if(X < x) x = X;
if(X > xmax) xmax = X;
if(Y < y) y = Y;
if(Y > ymax) ymax = Y;
}
}
}
++xmax; ++ymax;
w = xmax - x;
h = ymax - y;
}
struct Pix* po = pixCreate(w,h,32);
int ox, oy;
for(oy = 0, Y = y; Y < y + h; oy++, Y++) for(ox = 0, X = x; X < x + w; ox++, X++) {
if(((uint32_t*)p1b->data)[Y * W + X] != ((uint32_t*)p2b->data)[Y * W + X])
((uint32_t*)po->data)[oy * w + ox] = ((uint32_t*)p2b->data)[Y * W + X];
else
((uint32_t*)po->data)[oy * w + ox] = 0;
}
pixWritePng(out, po, 0);
dprintf(1, "wrote pixels x,y,w,h: %d,%d,%d,%d to %s\n", x,y,w,h,out);
return 0;
}
// Returns a full-resolution binary pix in which each cell over the given
// threshold is filled as a black square. pixDestroy after use.
// Edge cells, which have a zero 4-neighbour, are not marked.
Pix* IntGrid::ThresholdToPix(int threshold) const {
Pix* pix = pixCreate(tright().x() - bleft().x(),
tright().y() - bleft().y(), 1);
int cellsize = gridsize();
for (int y = 0; y < gridheight(); ++y) {
for (int x = 0; x < gridwidth(); ++x) {
if (GridCellValue(x, y) > threshold &&
GridCellValue(x - 1, y) > 0 && GridCellValue(x + 1, y) > 0 &&
GridCellValue(x, y - 1) > 0 && GridCellValue(x, y + 1) > 0) {
pixRasterop(pix, x * cellsize, tright().y() - ((y + 1) * cellsize),
cellsize, cellsize, PIX_SET, NULL, 0, 0);
}
}
}
return pix;
}
/*!
* pixInitAccumulate()
*
* Input: w, h (of accumulate array)
* offset (initialize the 32 bpp to have this
* value; not more than 0x40000000)
* Return: pixd (32 bpp), or null on error
*
* Notes:
* (1) The offset must be >= 0.
* (2) The offset is used so that we can do arithmetic
* with negative number results on l_uint32 data; it
* prevents the l_uint32 data from going negative.
* (3) Because we use l_int32 intermediate data results,
* these should never exceed the max of l_int32 (0x7fffffff).
* We do not permit the offset to be above 0x40000000,
* which is half way between 0 and the max of l_int32.
* (4) The same offset should be used for initialization,
* multiplication by a constant, and final extraction!
* (5) If you're only adding positive values, offset can be 0.
*/
PIX *
pixInitAccumulate(l_int32 w,
l_int32 h,
l_uint32 offset)
{
PIX *pixd;
PROCNAME("pixInitAccumulate");
if ((pixd = pixCreate(w, h, 32)) == NULL)
return (PIX *)ERROR_PTR("pixd not made", procName, NULL);
if (offset > 0x40000000)
offset = 0x40000000;
pixSetAllArbitrary(pixd, offset);
return pixd;
}
// Helper writes a grey image to a file for use by scrollviewer.
// Normally for speed we don't display the image in the layout debug windows.
// If textord_debug_images is true, we draw the image as a background to some
// of the debug windows. printable determines whether these
// images are optimized for printing instead of screen display.
static void WriteDebugBackgroundImage(bool printable, Pix* pix_binary) {
Pix* grey_pix = pixCreate(pixGetWidth(pix_binary),
pixGetHeight(pix_binary), 8);
// Printable images are light grey on white, but for screen display
// they are black on dark grey so the other colors show up well.
if (printable) {
pixSetAll(grey_pix);
pixSetMasked(grey_pix, pix_binary, 192);
} else {
pixSetAllArbitrary(grey_pix, 64);
pixSetMasked(grey_pix, pix_binary, 0);
}
AlignedBlob::IncrementDebugPix();
pixWrite(AlignedBlob::textord_debug_pix().string(), grey_pix, IFF_PNG);
pixDestroy(&grey_pix);
}
/*!
* pixaSplitPix()
*
* Input: pixs (with individual components on a lattice)
* nx (number of mosaic cells horizontally)
* ny (number of mosaic cells vertically)
* borderwidth (of added border on all sides)
* bordercolor (in our RGBA format: 0xrrggbbaa)
* Return: pixa, or null on error
*
* Notes:
* (1) This is a variant on pixaCreateFromPix(), where we
* simply divide the image up into (approximately) equal
* subunits. If you want the subimages to have essentially
* the same aspect ratio as the input pix, use nx = ny.
* (2) If borderwidth is 0, we ignore the input bordercolor and
* redefine it to white.
* (3) The bordercolor is always used to initialize each tiled pix,
* so that if the src is clipped, the unblitted part will
* be this color. This avoids 1 pixel wide black stripes at the
* left and lower edges.
*/
PIXA *
pixaSplitPix(PIX *pixs,
l_int32 nx,
l_int32 ny,
l_int32 borderwidth,
l_uint32 bordercolor)
{
l_int32 w, h, d, cellw, cellh, i, j;
PIX *pixt;
PIXA *pixa;
PROCNAME("pixaSplitPix");
if (!pixs)
return (PIXA *)ERROR_PTR("pixs not defined", procName, NULL);
if (nx <= 0 || ny <= 0)
return (PIXA *)ERROR_PTR("nx and ny must be > 0", procName, NULL);
borderwidth = L_MAX(0, borderwidth);
if ((pixa = pixaCreate(nx * ny)) == NULL)
return (PIXA *)ERROR_PTR("pixa not made", procName, NULL);
pixGetDimensions(pixs, &w, &h, &d);
cellw = (w + nx - 1) / nx; /* round up */
cellh = (h + ny - 1) / ny;
for (i = 0; i < ny; i++) {
for (j = 0; j < nx; j++) {
if ((pixt = pixCreate(cellw + 2 * borderwidth,
cellh + 2 * borderwidth, d)) == NULL)
return (PIXA *)ERROR_PTR("pixt not made", procName, NULL);
pixCopyColormap(pixt, pixs);
if (borderwidth == 0) { /* initialize full image to white */
if (d == 1)
pixClearAll(pixt);
else
pixSetAll(pixt);
}
else
pixSetAllArbitrary(pixt, bordercolor);
pixRasterop(pixt, borderwidth, borderwidth, cellw, cellh,
PIX_SRC, pixs, j * cellw, i * cellh);
pixaAddPix(pixa, pixt, L_INSERT);
}
}
return pixa;
}
/*!
* \brief pixConnCompIncrInit()
*
* \param[in] pixs 1 bpp
* \param[in] conn connectivity: 4 or 8
* \param[out] ppixd 32 bpp, with c.c. labelled
* \param[out] pptaa with pixel locations indexed by c.c.
* \param[out] pncc initial number of c.c.
* \return 0 if OK, 1 on error
*
* <pre>
* Notes:
* (1) This labels the connected components in a 1 bpp pix, and
* additionally sets up a ptaa that lists the locations of pixels
* in each of the components.
* (2) It can be used to initialize the output image and arrays for
* an application that maintains information about connected
* components incrementally as pixels are added.
* (3) pixs can be empty or have some foreground pixels.
* (4) The connectivity is stored in pixd->special.
* (5) Always initialize with the first pta in ptaa being empty
* and representing the background value (index 0) in the pix.
* </pre>
*/
l_int32
pixConnCompIncrInit(PIX *pixs,
l_int32 conn,
PIX **ppixd,
PTAA **pptaa,
l_int32 *pncc)
{
l_int32 empty, w, h, ncc;
PIX *pixd;
PTA *pta;
PTAA *ptaa;
PROCNAME("pixConnCompIncrInit");
if (ppixd) *ppixd = NULL;
if (pptaa) *pptaa = NULL;
if (pncc) *pncc = 0;
if (!ppixd || !pptaa || !pncc)
return ERROR_INT("&pixd, &ptaa, &ncc not all defined", procName, 1);
if (!pixs || pixGetDepth(pixs) != 1)
return ERROR_INT("pixs undefined or not 1 bpp", procName, 1);
if (conn != 4 && conn != 8)
return ERROR_INT("connectivity must be 4 or 8", procName, 1);
pixGetDimensions(pixs, &w, &h, NULL);
pixZero(pixs, &empty);
if (empty) {
*ppixd = pixCreate(w, h, 32);
pixSetSpp(*ppixd, 1);
pixSetSpecial(*ppixd, conn);
*pptaa = ptaaCreate(0);
pta = ptaCreate(1);
ptaaAddPta(*pptaa, pta, L_INSERT); /* reserve index 0 for background */
return 0;
}
/* Set up the initial labeled image and indexed pixel arrays */
if ((pixd = pixConnCompTransform(pixs, conn, 32)) == NULL)
return ERROR_INT("pixd not made", procName, 1);
pixSetSpecial(pixd, conn);
*ppixd = pixd;
if ((ptaa = ptaaIndexLabeledPixels(pixd, &ncc)) == NULL)
return ERROR_INT("ptaa not made", procName, 1);
*pptaa = ptaa;
*pncc = ncc;
return 0;
}
/*!
* \brief recogGetWindowedArea()
*
* \param[in] recog
* \param[in] index of template
* \param[in] x pixel position of left hand edge of template
* \param[out] pdely y shift of template relative to pix1
* \param[out] pwsum number of fg pixels in window of pixs
* \return 0 if OK, 1 on error
*
* <pre>
* Notes:
* (1) This is called after the best path has been found through
* the trellis, in order to produce a correlation that can be used
* to evaluate the confidence we have in the identification.
* The correlation is |1 \& 2|^2 / (|1| * |2|).
* |1 \& 2| is given by the count array, |2| is found from
* nasum_u[], and |1| is wsum returned from this function.
* </pre>
*/
static l_int32
recogGetWindowedArea(L_RECOG *recog,
l_int32 index,
l_int32 x,
l_int32 *pdely,
l_int32 *pwsum)
{
l_int32 w1, h1, w2, h2;
PIX *pix1, *pix2, *pixt;
L_RDID *did;
PROCNAME("recogGetWindowedArea");
if (pdely) *pdely = 0;
if (pwsum) *pwsum = 0;
if (!pdely || !pwsum)
return ERROR_INT("&dely and &wsum not both defined", procName, 1);
if (!recog)
return ERROR_INT("recog not defined", procName, 1);
if ((did = recogGetDid(recog)) == NULL)
return ERROR_INT("did not defined", procName, 1);
if (index < 0 || index >= did->narray)
return ERROR_INT("invalid index", procName, 1);
pix1 = did->pixs;
pixGetDimensions(pix1, &w1, &h1, NULL);
if (x >= w1)
return ERROR_INT("invalid x position", procName, 1);
pix2 = pixaGetPix(recog->pixa_u, index, L_CLONE);
pixGetDimensions(pix2, &w2, &h2, NULL);
if (w1 < w2) {
L_INFO("template %d too small\n", procName, index);
pixDestroy(&pix2);
return 0;
}
*pdely = did->delya[index][x];
pixt = pixCreate(w2, h1, 1);
pixRasterop(pixt, 0, *pdely, w2, h2, PIX_SRC, pix2, 0, 0);
pixRasterop(pixt, 0, 0, w2, h1, PIX_SRC & PIX_DST, pix1, x, 0);
pixCountPixels(pixt, pwsum, recog->sumtab);
pixDestroy(&pix2);
pixDestroy(&pixt);
return 0;
}
/*!
* \brief pixReadMemWebP()
*
* \param[in] filedata webp compressed data in memory
* \param[in] filesize number of bytes in data
* \return pix 32 bpp, or NULL on error
*
* <pre>
* Notes:
* (1) When the encoded data only has 3 channels (no alpha),
* WebPDecodeRGBAInto() generates a raster of 32-bit pixels, with
* the alpha channel set to opaque (255).
* (2) We don't need to use the gnu runtime functions like fmemopen()
* for redirecting data from a stream to memory, because
* the webp library has been written with memory-to-memory
* functions at the lowest level (which is good!). And, in
* any event, fmemopen() doesn't work with l_binaryReadStream().
* </pre>
*/
PIX *
pixReadMemWebP(const l_uint8 *filedata,
size_t filesize)
{
l_uint8 *out = NULL;
l_int32 w, h, has_alpha, wpl, stride;
l_uint32 *data;
size_t size;
PIX *pix;
WebPBitstreamFeatures features;
PROCNAME("pixReadMemWebP");
if (!filedata)
return (PIX *)ERROR_PTR("filedata not defined", procName, NULL);
if (WebPGetFeatures(filedata, filesize, &features))
return (PIX *)ERROR_PTR("Invalid WebP file", procName, NULL);
w = features.width;
h = features.height;
has_alpha = features.has_alpha;
/* Write from compressed Y,U,V arrays to pix raster data */
pix = pixCreate(w, h, 32);
pixSetInputFormat(pix, IFF_WEBP);
if (has_alpha) pixSetSpp(pix, 4);
data = pixGetData(pix);
wpl = pixGetWpl(pix);
stride = wpl * 4;
size = stride * h;
out = WebPDecodeRGBAInto(filedata, filesize, (uint8_t *)data, size,
stride);
if (out == NULL) { /* error: out should also point to data */
pixDestroy(&pix);
return (PIX *)ERROR_PTR("WebP decode failed", procName, NULL);
}
/* The WebP API expects data in RGBA order. The pix stores
* in host-dependent order with R as the MSB and A as the LSB.
* On little-endian machines, the bytes in the word must
* be swapped; e.g., R goes from byte 0 (LSB) to byte 3 (MSB).
* No swapping is necessary for big-endians. */
pixEndianByteSwap(pix);
return pix;
}