def text_line_segmentation(binary, scale=None, gray=None, num_col = 1):
"""Given a binary image, compute a complete segmentation into
lines, computing both columns and text lines."""
binary = array(binary, 'B')
if scale is None:
scale = psegutils.estimate_scale(binary)
# do the column finding
if num_col > 1:
colseps, binary = compute_colseps(binary, scale)
else:
colseps = np.zeros(binary.shape)
# now compute the text line seeds
bottom, top, boxmap = compute_gradmaps(binary, scale)
seeds = compute_line_seeds(binary, bottom, top, colseps, scale)
# spread the text line seeds to all the remaining components
llabels = morph.propagate_labels(boxmap, seeds, conflict=0)
spread = morph.spread_labels(seeds, maxdist=scale)
llabels = where(llabels > 0, llabels, spread * binary)
segmentation = llabels * binary
lines = psegutils.compute_lines(segmentation, scale, 0.8)
line_ims = []
for l in lines:
if gray is None:
binline = psegutils.extract_masked(1-binary, l, pad=0)
else:
binline = psegutils.extract_masked(gray, l, pad=0)
binline = pad_by(binline, 10, invert=False)
line_ims.append(binline)
return line_ims, lines
def extract(image):
try:
binary = ocrolib.read_image_binary(image)
binary = 1-binary
scale = psegutils.estimate_scale(binary)
segmentation = compute_segmentation(binary,scale)
# ...lines = compute_lines(segmentation,scale)
# compute the reading order
lines = psegutils.compute_lines(segmentation,scale)
order = psegutils.reading_order([l.bounds for l in lines])
lsort = psegutils.topsort(order)
# renumber the labels so that they conform to the specs
nlabels = amax(compute_segmentation)+1
renumber = zeros(nlabels,'i')
for i,v in enumerate(lsort): renumber[lines[v].label] = 0x010000+(i+1)
segmentation = renumber[segmentation]
outputdir = "http://127.0.0.1:5000/uploads/"
lines = [lines[i] for i in lsort]
ocrolib.write_page_segmentation("%s.pseg.png"%outputdir,segmentation)
cleaned = ocrolib.remove_noise(binary,args.noise)
for i,l in enumerate(lines):
binline = psegutils.extract_masked(1-cleaned,l,pad=args.pad,expand=args.expand)
ocrolib.write_image_binary("%s/01%04x.bin.png"%(outputdir,i+1),binline)
#print "%6d"%i,fname,"%4.1f"%scale,len(lines)
except:
print ('error')
def calc(self, objects, scale):
if self.binpage is None:
return
tt = time()
bottom, top, boxmap = compute_gradmaps(self.binpage, scale)
# DSHOW('hihi', [0.5*bottom+0.5*top,self.binpage, boxmap])
seeds0 = compute_line_seeds(self.binpage, bottom, top, scale)
seeds, _ = morph.label(seeds0)
llabels = morph.propagate_labels(boxmap, seeds, conflict=0)
spread = spread_labels(seeds, maxdist=scale)
llabels = where(llabels > 0, llabels, spread * self.binpage)
segmentation = llabels * self.binpage
self.binpage = ocrolib.remove_noise(self.binpage, args.noise)
lines = psegutils.compute_lines(segmentation, scale)
binpage_reversed = 1 - self.binpage
# print 'pre line ', time() - tt
tt = time()
self.lines = []
for i, l in enumerate(lines):
tt = time()
binline = psegutils.extract_masked(
binpage_reversed, l, pad=args.pad,
expand=args.expand) # black text
binline = (1 - binline)
le = lineest.CenterNormalizer(binline.shape[0]) # white text
binline = binline.astype(float)
le.measure(binline)
binline = le.normalize(binline)
binline = where(binline > 0.5, 0, 1) # black text
# print 'line time ', time()-tt
print '-----------------------'
pilimg = Image.fromarray((binline * 255).astype(uint8))
pred_legacy = pytesseract.image_to_string(pilimg,
lang='eng',
config='--oem 0 --psm 7')
print '00', pred_legacy
pred_lstm = pytesseract.image_to_string(pilimg,
lang='eng',
config='--oem 1 --psm 7')
print '11', pred_lstm
# ASHOW('line',binline, scale=2.0)
## pred_both = pytesseract.image_to_string(pilimg,lang='vie', config='--oem 2 --psm 7')
## print '22', pred_both
result = psegutils.record(bounds=l.bounds,
text1=pred_legacy,
text2=pred_lstm,
img=binline)
self.lines.append(result)
def printResult(self, outputfile):
# Some pre-process
# print 'text area before'
# cv2.imshow('patch', self.patch)
# cv2.waitKey(-1)
if self.name == 'CMND cu - 9 so':
k = 0.45
else:
k = 0.33
patch = sharpen(self.patch)
binary = sauvola(patch,
w=int(self.template.shape[1] / 24.5 * 2),
k=k,
scaledown=0.5,
reverse=True)
binary = cv2.bitwise_and(binary, binary, mask=self.patch_mask)
# print 'text area after'
# cv2.imshow('patch', binary*255)
# cv2.waitKey(-1)
dotremoved = binary
scale = self.scale
# Line extraction copied from Ocropus source code
bottom, top, boxmap = compute_gradmaps(dotremoved, scale)
seeds0 = compute_line_seeds(dotremoved, bottom, top, scale)
seeds, _ = morph.label(seeds0)
llabels = morph.propagate_labels(boxmap, seeds, conflict=0)
spread = spread_labels(seeds, maxdist=scale)
llabels = where(llabels > 0, llabels, spread * dotremoved)
segmentation = llabels * dotremoved
dotremoved = ocrolib.remove_noise(dotremoved, 8)
lines = psegutils.compute_lines(segmentation, scale / 2)
binpage_reversed = 1 - dotremoved
self.lines = []
readrs = dict.fromkeys(self.linepos1.keys(), u'')
lines = sorted(lines, key=lambda x: x.bounds[1].start)
for i, l in enumerate(lines):
# Line extraction copied from Ocropus source code
binline = psegutils.extract_masked(binpage_reversed,
l,
pad=int(scale / 2),
expand=0) # black text
binline = (1 - binline)
le = lineest.CenterNormalizer(binline.shape[0]) # white text
binline = binline.astype(float)
le.measure(binline)
binline = le.normalize(binline)
# print 'normalized'
# cv2.imshow('line', binline)
# cv2.waitKey(-1)
binline = cv2.resize(binline, None, fx=2.0, fy=2.0)
# print 'resized'
# cv2.imshow('line', binline)
# cv2.waitKey(-1)
binline = where(binline > 0.5, uint8(0), uint8(255)) # black text
# print 'black text'
# cv2.imshow('line', binline)
# cv2.waitKey(-1)
# pilimg = Image.fromarray(binline)
pos = l.bounds[0].stop
left = (l.bounds[1].start < self.template.shape[1] / 2)
# Prediction using Tesseract 4.0
if pos > self.linepos1['idNumber'][0] and pos < self.linepos1[
'idNumber'][1]: #ID, all numbers
pred = ocr(
binline,
config=
'--oem 0 --psm 7 -c tessedit_char_whitelist=0123456789')
readrs['idNumber'] += pred + u' '
elif pos > self.linepos1['dateOfBirth'][0] and pos < self.linepos1[
'dateOfBirth'][1]: # DOB, number, - , /
pred = ocr(
binline,
config=
'--oem 1 --psm 7 -c tessedit_char_whitelist=0123456789-/')
readrs['dateOfBirth'] += pred + u' '
elif left and pos > self.linepos1['Gender'][
0] and pos < self.linepos1['Gender'][1]:
pred = ocr(binline, config='--oem 1 --psm 7 -l vie')
readrs['Gender'] += pred + u' '
elif (not left) and pos > self.linepos1['Dantoc'][
0] and pos < self.linepos1['Dantoc'][1]:
pred = ocr(binline, config='--oem 1 --psm 7 -l vie')
readrs['Dantoc'] += pred + u' '
elif pos > self.linepos1['NguyenQuan'][0] and pos < self.linepos1[
'NguyenQuan'][1]:
pred = ocr(binline, config='--oem 1 --psm 7 -l vie')
readrs['NguyenQuan'] += pred + u' '
elif pos > self.linepos1['fullName'][0] and pos < self.linepos1[
'fullName'][1]:
pred = ocr(binline, config='--oem 1 --psm 7 -l vie')
readrs['fullName'] += pred + u' '
# else:
# pred = ocr(binline, config='--oem 1 --psm 7 -l vie')
# print 'unknown ', unicode2ascii(pred), 'y:', l.bounds[0], 'x:', l.bounds[1]
for k in readrs:
readrs[k] = (readrs[k].replace(u'²',
u'2').replace(u'º', u'o').replace(
u'»', u'-')).strip()
if len(readrs[k]) == 0:
readrs[k] = None
if self.name == 'CMND moi - 12 so':
readrs['type'] = 'CMND Mới - 12 Số'
elif self.name == 'Can Cuoc Cong Dan':
readrs['type'] = 'Căn Cước Công Dân'
elif self.name == 'CMND cu - 9 so':
readrs['type'] = 'CMND Cũ - 9 Số'
readrs['NgayHetHan'] = None
outputfile.write(json.dumps(readrs))
def analyze_page_layout(binary, gray, rgb=None):
hscale = 1.0 # Non-standard scaling of horizontal parameters.
vscale = 1.0 # Non-standard scaling of vertical parameters.
threshold = 0.2 # baseline threshold.
usegauss = True # Use gaussian instead of uniform.
maxseps = 0 # Maximum black column separators.
sepwiden = 10 # Widen black separators (to account for warping).
blackseps = True
maxcolseps = 3 # Maximum # whitespace column separators.
csminheight = 10 # Minimum column height (units=scale).
noise = 8 # Noise threshold for removing small components from lines.
gray_output = True # Output grayscale lines as well, which are extracted from the grayscale version of the pages.
pad = 3 # Padding for extracted lines.
expand = 3 # Expand mask for grayscale extraction.
if False:
bin_image_filepath = './ocropy_test.bin.png'
gray_image_filepath = './ocropy_test.nrm.png'
binary = ocrolib.read_image_binary(bin_image_filepath)
gray = ocrolib.read_image_gray(gray_image_filepath)
binary = 1 - binary # Invert.
scale = psegutils.estimate_scale(binary)
segmentation = compute_segmentation(binary,
scale,
blackseps,
maxseps,
maxcolseps,
csminheight,
sepwiden,
usegauss,
vscale,
hscale,
threshold,
quiet=True)
lines = psegutils.compute_lines(segmentation, scale)
order = psegutils.reading_order([l.bounds for l in lines])
lsort = psegutils.topsort(order)
# Renumber the labels so that they conform to the specs.
nlabels = np.amax(segmentation) + 1
renumber = np.zeros(nlabels, 'i')
for i, v in enumerate(lsort):
renumber[lines[v].label] = 0x010000 + (i + 1)
segmentation = renumber[segmentation] # Image.
lines = [lines[i] for i in lsort]
# Visualize bounding boxes.
if False:
if rgb is not None:
# REF [function] >> extract_masked() in ${OCROPY_HOME}/ocrolib/psegutils.py.
for l in lines:
y0, x0, y1, x1 = [
int(x) for x in [
l.bounds[0].start, l.bounds[1].start, l.bounds[0].stop,
l.bounds[1].stop
]
]
cv2.rectangle(rgb, (x0, y0), (x1, y1), (0, 0, 255), 1,
cv2.LINE_AA)
cv2.imshow('Image', rgb)
cv2.waitKey(0)
# Output everything.
if False:
if not os.path.exists(outputdir):
os.mkdir(outputdir)
ocrolib.write_page_segmentation("%s.pseg.png" % outputdir,
segmentation)
cleaned = ocrolib.remove_noise(binary, noise)
for i, l in enumerate(lines):
binline = psegutils.extract_masked(1 - cleaned,
l,
pad=pad,
expand=expand) # Image.
ocrolib.write_image_binary(
"%s/01%04x.bin.png" % (outputdir, i + 1), binline)
if gray_output:
grayline = psegutils.extract_masked(gray,
l,
pad=pad,
expand=expand) # Image.
ocrolib.write_image_gray(
"%s/01%04x.nrm.png" % (outputdir, i + 1), grayline)
def processPngFile(outRoot, origFile, fileNum):
baseName = os.path.basename(origFile)
baseBase, _ = os.path.splitext(baseName)
outDir = os.path.join(outRoot, "%s.%03d" % (baseBase, fileNum))
inFile = os.path.join(outDir, baseName)
os.makedirs(outDir, exist_ok=True)
shutil.copy(origFile, inFile)
inBase, _ = ocrolib.allsplitext(inFile)
print("** inBase=%s" % inBase)
# print("** binBase=%s" % binBase)
fname = inFile
outputdir = inBase
binFile = inBase + ".bin.png"
outFile = inBase + ".out.png"
outRoot2, outDir2 = os.path.split(outRoot)
outFile2 = os.path.join(outRoot2, "%s.out" % outDir2, baseName)
print("outFile2=%s" % outFile2)
# assert False
grayFile = inBase + ".nrm.png"
psegFile = inBase + ".pseg.png"
print(" inFile=%s" % inFile)
print(" binFile=%s" % binFile)
print("grayFile=%s" % grayFile)
print(" outFile=%s" % outFile)
assert inFile and binFile
assert outFile != inFile
assert outFile != binFile
if not binarize(inFile, binFile, grayFile):
binExists = os.path.exists(binFile)
print("Couldn't binarize inFile=%s binFile=%s exists=%s" %
(inFile, binFile, binExists))
return False
binary = ocrolib.read_image_binary(binFile)
print("$$ %s=%s" % (binFile, desc(binary)))
height, width = binary.shape
checktype(binary, ABINARY2)
check = check_page(np.amax(binary) - binary)
if check is not None:
print("%s SKIPPED %s (use -n to disable this check)" % (inFile, check))
return False
# if args.gray:
# if os.path.exists(base+".nrm.png"):
# gray = ocrolib.read_image_gray(base+".nrm.png")
# checktype(gray, GRAYSCALE)
# else:
# print_error("Grayscale version %s.nrm.png not found. Use ocropus-nlbin for creating " +
# "normalized grayscale version of the pages as well." % base)
# return
binary = 1 - binary # invert
scale = psegutils.estimate_scale(binary)
print("scale %f" % scale)
if np.isnan(scale) or scale > 1000.0:
print("%s: bad scale (%g); skipping\n" % (fname, scale))
return False
# find columns and text lines
print("computing segmentation")
segmentation = compute_segmentation(binary, scale)
if np.amax(segmentation) > maxlines:
print("%s: too many lines %g" % (fname, np.amax(segmentation)))
return False
print("segmentation=%s" % desc(segmentation))
print("number of lines %g" % np.amax(segmentation))
# compute the reading order
print("finding reading order")
lines = psegutils.compute_lines(segmentation, scale)
order = psegutils.reading_order([l.bounds for l in lines])
lsort = psegutils.topsort(order)
print("$$ lsort = %d = %s...%s" % (len(lsort), lsort[:10], lsort[-10:]))
# renumber the labels so that they conform to the specs
nlabels = np.amax(segmentation) + 1
renumber = np.zeros(nlabels, 'i')
for i, v in enumerate(lsort):
renumber[lines[v].label] = 0x010000 + (i + 1)
segmentation = renumber[segmentation]
# finally, output everything
print("writing lines")
if not os.path.exists(outputdir):
os.mkdir(outputdir)
lines = [lines[i] for i in lsort]
ocrolib.write_page_segmentation("%s.pseg.png" % outputdir, segmentation)
cleaned = ocrolib.remove_noise(binary, noise)
for i, l in enumerate(lines):
binline = psegutils.extract_masked(1 - cleaned,
l,
pad=pad,
expand=expand)
ocrolib.write_image_binary("%s/01%04x.bin.png" % (outputdir, i + 1),
binline)
# if args.gray:
# grayline = psegutils.extract_masked(
# gray, l, pad=args.pad, expand=args.expand)
# ocrolib.write_image_gray("%s/01%04x.nrm.png" % (outputdir, i+1), grayline)
print("%6d %s %4.1f %d" % (i, fname, scale, len(lines)))
# to proceed, we need a pseg file and a subdirectory containing text lines
assert os.path.exists(psegFile), "%s: no such file" % psegFile
assert os.path.isdir(inBase), "%s: no such directory" % inBase
# iterate through the text lines in reading order, based on the page segmentation file
pseg = ocrolib.read_page_segmentation(psegFile)
print("$$ %s=%s" % (psegFile, desc(pseg)))
regions = ocrolib.RegionExtractor()
print("$$ regions=%s" % regions)
regions.setPageLines(pseg)
im = Image.open(inFile)
print("~~%s %s" % (inFile, im.size))
print("$$ regions=%s=%s" % (regions, sorted(regions.__dict__)))
print("$$ regions.length=%s" % regions.length())
n = regions.length()
for i in range(1, n):
id = regions.id(i)
y0, x0, y1, x1 = regions.bbox(i)
# print("%5d: 0x%05X %s %d x %d" %
# (i, id, [y0, x0, y1, x1], y1 - y0, x1 - x0))
draw = ImageDraw.Draw(im)
draw.rectangle((x0, y0, x1, y1), outline=(255, 0, 0), width=3)
draw.rectangle((x0, y0, x1, y1), outline=(0, 0, 255), width=0)
# draw.rectangle((x0, y0, x1, y1), outline=255, width=5)
# draw.rectangle((x0, y0, x1, y1), outline=10, width=1)
del draw
# write output files
print("outFile=%s" % outFile)
im.save(outFile, "PNG")
print("outFile2=%s" % outFile2)
outDir2 = os.path.dirname(outFile2)
os.makedirs(outDir2, exist_ok=True)
im.save(outFile2, "PNG")
assert os.path.exists(outFile2)
# outFile3, _ = os.path.splitext(outFile)
# outFile3 = "%s.jpg" % outFile3
# print("outFile3=%s" % outFile3)
# im.save(outFile3, "JPEG")
# assert os.path.exists(outFile3)
return True
def process(job):
imagepath, i = job
global base
base, _ = ocrolib.allsplitext(imagepath)
outputdir = base
imagename_base = os.path.basename(os.path.normpath(base))
try:
binary = ocrolib.read_image_binary(imagepath)
except IOError:
if ocrolib.trace: traceback.print_exc()
print_error("cannot open either %s.bin.png or %s" % (base, imagepath))
return
checktype(binary, ABINARY2)
if not args['nocheck']:
check = check_page(amax(binary) - binary)
if check is not None:
print_error("%s SKIPPED %s (use -n to disable this check)" %
(imagepath, check))
return
binary = 1 - binary # invert
if args['scale'] == 0:
scale = psegutils.estimate_scale(binary)
else:
scale = args['scale']
print_info("scale %f" % (scale))
if isnan(scale) or scale > 1000.0:
print_error("%s: bad scale (%g); skipping\n" % (imagepath, scale))
return
if scale < args['minscale']:
print_error("%s: scale (%g) less than --minscale; skipping\n" %
(imagepath, scale))
return
# find columns and text lines
if not args['quiet']: print_info("computing segmentation")
segmentation = compute_segmentation(binary, scale)
if amax(segmentation) > args['maxlines']:
print_error("%s: too many lines %g" % (imagepath, amax(segmentation)))
return
if not args['quiet']: print_info("number of lines %g" % amax(segmentation))
# compute the reading order
if not args['quiet']: print_info("finding reading order")
lines = psegutils.compute_lines(segmentation, scale)
order = psegutils.reading_order([l.bounds for l in lines])
lsort = psegutils.topsort(order)
# renumber the labels so that they conform to the specs
nlabels = amax(segmentation) + 1
renumber = zeros(nlabels, 'i')
for i, v in enumerate(lsort):
renumber[lines[v].label] = 0x010000 + (i + 1)
segmentation = renumber[segmentation]
# finally, output everything
if not args['quiet']: print_info("writing lines")
if not os.path.exists(outputdir):
os.mkdir(outputdir)
lines = [lines[i] for i in lsort]
ocrolib.write_page_segmentation("%s.pseg.png" % outputdir, segmentation)
cleaned = ocrolib.remove_noise(binary, args['noise'])
for i, l in enumerate(lines):
binline = psegutils.extract_masked(1 - cleaned,
l,
pad=args['pad'],
expand=args['expand'])
ocrolib.write_image_binary(
"%s/%s_01%04x.bin.png" % (outputdir, imagename_base, i + 1),
binline)
print_info("%6d %s %4.1f %d" % (i, imagepath, scale, len(lines)))
return outputdir
def process1(job):
fname, i = job
global base
base, _ = ocrolib.allsplitext(fname)
outputdir = base
try:
binary = ocrolib.read_image_binary(base + ".bin.png")
except IOError:
try:
binary = ocrolib.read_image_binary(fname)
except IOError:
if ocrolib.trace:
traceback.print_exc()
print("cannot open either", base + ".bin.png", "or", fname)
return
checktype(binary, ABINARY2)
if not args.nocheck:
check = check_page(amax(binary) - binary)
if check is not None:
print(fname, "SKIPPED", check, "(use -n to disable this check)")
return
if args.gray:
if os.path.exists(base + ".nrm.png"):
gray = ocrolib.read_image_gray(base + ".nrm.png")
checktype(gray, GRAYSCALE)
binary = 1 - binary # invert
if args.scale == 0:
scale = psegutils.estimate_scale(binary)
else:
scale = args.scale
print("scale", scale)
if isnan(scale) or scale > 1000.0:
sys.stderr.write("%s: bad scale (%g); skipping\n" % (fname, scale))
return
if scale < args.minscale:
sys.stderr.write("%s: scale (%g) less than --minscale; skipping\n" %
(fname, scale))
return
# find columns and text lines
if not args.quiet:
print("computing segmentation")
segmentation = compute_segmentation(binary, scale)
if amax(segmentation) > args.maxlines:
print(fname, ": too many lines", amax(segmentation))
return
if not args.quiet:
print("number of lines", amax(segmentation))
# compute the reading order
if not args.quiet:
print("finding reading order")
lines = psegutils.compute_lines(segmentation, scale)
order = psegutils.reading_order([l.bounds for l in lines])
lsort = psegutils.topsort(order)
# renumber the labels so that they conform to the specs
nlabels = amax(segmentation) + 1
renumber = zeros(nlabels, 'i')
for i, v in enumerate(lsort):
renumber[lines[v].label] = 0x010000 + (i + 1)
segmentation = renumber[segmentation]
# finally, output everything
if not args.quiet:
print("writing lines")
if not os.path.exists(outputdir):
os.mkdir(outputdir)
lines = [lines[i] for i in lsort]
ocrolib.write_page_segmentation("%s.pseg.png" % outputdir, segmentation)
cleaned = ocrolib.remove_noise(binary, args.noise)
for i, l in enumerate(lines):
binline = psegutils.extract_masked(1 - cleaned,
l,
pad=args.pad,
expand=args.expand)
ocrolib.write_image_binary("%s/01%04x.bin.png" % (outputdir, i + 1),
binline)
if args.gray:
grayline = psegutils.extract_masked(gray,
l,
pad=args.pad,
expand=args.expand)
ocrolib.write_image_gray("%s/01%04x.nrm.png" % (outputdir, i + 1),
grayline)
print("%6d" % i, fname, "%4.1f" % scale, len(lines))