def setLine(self,image,cseg=None):
# print "WS",image.shape,cseg.shape
# ocrolib.write_image_gray("_wmodel.png",image)
assert mean(image)<0.5*amax(image)
if amax(image)<1e-6: raise BadImage()
self.image = array(image*(1.0/amax(image)),'f')
if cseg is None:
if "segmenter" not in dir(self):
self.segmenter = lineseg.DPSegmentLine()
cseg = self.segmenter.charseg(amax(image)-image)
morph.showlabels(cseg)
self.cseg = cseg
self.line_params = lineproc.seg_geometry(cseg,math=0)
def setLine(self, image, cseg=None):
# print "WS",image.shape,cseg.shape
# ocrolib.write_image_gray("_wmodel.png",image)
assert mean(image) < 0.5 * amax(image)
if amax(image) < 1e-6: raise BadImage()
self.image = array(image * (1.0 / amax(image)), 'f')
if cseg is None:
if "segmenter" not in dir(self):
self.segmenter = lineseg.DPSegmentLine()
cseg = self.segmenter.charseg(amax(image) - image)
morph.showlabels(cseg)
self.cseg = cseg
self.line_params = lineproc.seg_geometry(cseg, math=0)
def compute_geomaps(fnames,shapedict,old_model,use_gt=1,size=32,debug=0,old_order=1):
"""Given a shape dictionary and an existing line geometry
estimator, compute updated geometric maps for each entry
in the shape dictionary."""
if debug>0: gray(); ion()
shape = (shapedict.k,size,size)
bls = zeros(shape)
xls = zeros(shape)
count = 0
for fno,fname in enumerate(fnames):
if fno%20==0: print fno,fname,count
if use_gt:
# don't use lines with many capital letters for training because
# they result in bad models
gt = ocrolib.read_text(ocrolib.fvariant(fname,"txt","gt"))
if len(re.sub(r'[^A-Z]','',gt))>=0.3*len(re.sub(r'[^a-z]','',gt)): continue
if len(re.sub(r'[^0-9]','',gt))>=0.3*len(re.sub(r'[^a-z]','',gt)): continue
image = 1-ocrolib.read_image_gray(fname)
if debug>0 and fno%debug==0: clf(); subplot(411); imshow(image)
try:
blp,xlp = old_model.lineFit(image,order=old_order)
except:
traceback.print_exc()
continue
blimage = zeros(image.shape)
h,w = image.shape
for x in range(w): blimage[clip(int(polyval(blp,x)),0,h-1),x] = 1
xlimage = zeros(image.shape)
for x in range(w): xlimage[clip(int(polyval(xlp,x)),0,h-1),x] = 1
if debug>0 and fno%debug==0:
subplot(413); imshow(xlimage+0.3*image)
subplot(414); imshow(blimage+0.3*image)
try:
seg = lineseg.ccslineseg(image)
except:
continue
if debug>0 and fno%debug==0: subplot(412); morph.showlabels(seg)
shape = None
for sub,transform,itransform_add in extract_chars(seg):
if shape is None: shape = sub.shape
assert sub.shape==shape
count += 1
best = shapedict.predict1(sub)
bls[best] += transform(blimage)
xls[best] += transform(xlimage)
if debug==1: ginput(1,100)
elif debug>1: ginput(1,0.01)
for i in range(len(bls)): bls[i] *= bls[i].shape[1]*1.0/max(1e-6,sum(bls[i]))
for i in range(len(xls)): xls[i] *= xls[i].shape[1]*1.0/max(1e-6,sum(xls[i]))
return bls,xls
def charseg(self, line):
"""Segment a text line into potential character parts."""
assert mean(line) > 0.5 * amax(line)
line0 = amax(line) - line
ccuts = contourcuts(line0,
maxdist=self.maxdist,
minrange=self.minrange,
mincdist=self.mincdist,
sigma=self.sigma,
r=self.rr,
s=self.s)
line = maximum(0, line0 - ccuts)
# line = line+self.ledge*maximum(0,roll(line,-1,1)-line)
tracks = dplineseg2(line,
imweight=self.imweight,
bweight=self.bweight,
diagweight=self.diagweight,
debug=self.debug,
r=self.r)
tracks = array(tracks < 0.5 * amax(tracks), 'i')
tracks, _ = morph.label(tracks)
self.tracks = tracks
stracks = morph.spread_labels(tracks)
rsegs = stracks * (line0 > 0.5 * amax(line0))
if self.debug:
figure("temp")
subplot(311)
morph.showlabels(tracks)
subplot(312)
morph.showlabels(stracks)
subplot(313)
morph.showlabels(rsegs)
raw_input()
return morph.renumber_by_xcenter(rsegs)
def charseg(self, line):
"""Segment a text line into potential character parts."""
assert mean(line) > 0.5 * amax(line)
line = amax(line) - line
# line = line+self.ledge*maximum(0,roll(line,-1,1)-line)
tracks = dplineseg2(line,
imweight=self.imweight,
bweight=self.bweight,
diagweight=self.diagweight,
debug=self.debug,
r=self.r,
threshold=self.threshold)
tracks = array(tracks < 0.5 * amax(tracks), 'i')
tracks, _ = morph.label(tracks)
self.tracks = tracks # for debuggin
stracks = morph.spread_labels(tracks)
rsegs = stracks * (line > 0.5 * amax(line))
if 0:
figure("temp")
subplot(311)
morph.showlabels(tracks)
subplot(312)
morph.showlabels(stracks)
subplot(313)
morph.showlabels(rsegs)
raw_input()
return morph.renumber_by_xcenter(rsegs)
def charseg(self,line):
"""Segment a text line into potential character parts."""
assert mean(line)>0.5*amax(line)
line = amax(line)-line
# line = line+self.ledge*maximum(0,roll(line,-1,1)-line)
tracks = dplineseg2(line,imweight=self.imweight,bweight=self.bweight,
diagweight=self.diagweight,debug=self.debug,r=self.r,
threshold=self.threshold)
tracks = array(tracks<0.5*amax(tracks),'i')
tracks,_ = morph.label(tracks)
self.tracks = tracks # for debuggin
stracks = morph.spread_labels(tracks)
rsegs = stracks*(line>0.5*amax(line))
if 0:
figure("temp")
subplot(311); morph.showlabels(tracks)
subplot(312); morph.showlabels(stracks)
subplot(313); morph.showlabels(rsegs)
raw_input()
return morph.renumber_by_xcenter(rsegs)
def charseg(self,line):
"""Segment a text line into potential character parts."""
assert mean(line)>0.5*amax(line)
line0 = amax(line)-line
ccuts = contourcuts(line0,maxdist=self.maxdist,minrange=self.minrange,
mincdist=self.mincdist,sigma=self.sigma,r=self.rr,s=self.s)
line = maximum(0,line0-ccuts)
# line = line+self.ledge*maximum(0,roll(line,-1,1)-line)
tracks = dplineseg2(line,imweight=self.imweight,bweight=self.bweight,
diagweight=self.diagweight,debug=self.debug,r=self.r)
tracks = array(tracks<0.5*amax(tracks),'i')
tracks,_ = morph.label(tracks)
self.tracks = tracks
stracks = morph.spread_labels(tracks)
rsegs = stracks*(line0>0.5*amax(line0))
if self.debug:
figure("temp")
subplot(311); morph.showlabels(tracks)
subplot(312); morph.showlabels(stracks)
subplot(313); morph.showlabels(rsegs)
raw_input()
return morph.renumber_by_xcenter(rsegs)
subparsers = parser.add_subparsers(dest="subcommand")
test = subparsers.add_parser("test")
test.add_argument("--imweight",type=float,default=4,help="image weight (%(default)f")
test.add_argument("--bweight",type=float,default=-1,help="left border weight (%(default)f)")
test.add_argument("--diagweight",type=float,default=1,help="additional diagonal weight (%(default)f)")
test.add_argument("--r",type=int,default=1,help="range for diagonal steps (%(default)d)")
test.add_argument("--threshold",type=float,default=0.5)
test.add_argument("files",nargs="+",default=[])
# test2 = subparsers.add_parser("test2")
args = parser.parse_args()
if args.subcommand=="test":
segmenter = DPSegmentLine(imweight=args.imweight,
bweight=args.bweight,
diagweight=args.diagweight,
r=args.r,
threshold=args.threshold,
debug=1)
ion(); gray()
for fname in args.files:
print fname
image = ocrolib.read_image_gray(fname)
segmentation = segmenter.charseg(image)
figure("output")
subplot(211); imshow(image)
subplot(212); morph.showlabels(segmentation)
raw_input()
else:
parser.print_help()
sys.exit(0)
subparsers = parser.add_subparsers(dest="subcommand")
test = subparsers.add_parser("test")
test.add_argument("--imweight",type=float,default=4,help="image weight (%(default)f")
test.add_argument("--bweight",type=float,default=-1,help="left border weight (%(default)f)")
test.add_argument("--diagweight",type=float,default=1,help="additional diagonal weight (%(default)f)")
test.add_argument("--r",type=int,default=1,help="range for diagonal steps (%(default)d)")
test.add_argument("--threshold",type=float,default=0.5)
test.add_argument("files",nargs="+",default=[])
# test2 = subparsers.add_parser("test2")
args = parser.parse_args()
if args.subcommand=="test":
segmenter = DPSegmentLine(imweight=args.imweight,
bweight=args.bweight,
diagweight=args.diagweight,
r=args.r,
threshold=args.threshold,
debug=1)
ion(); gray()
for fname in args.files:
print fname
image = ocrolib.read_image_gray(fname)
segmentation = segmenter.charseg(image)
figure("output")
subplot(211); imshow(image)
subplot(212); morph.showlabels(segmentation)
raw_input()
else:
parser.print_help()
sys.exit(0)
test.add_argument("--r",
type=int,
default=1,
help="range for diagonal steps (%(default)d)")
test.add_argument("--threshold", type=float, default=0.5)
test.add_argument("files", nargs="+", default=[])
# test2 = subparsers.add_parser("test2")
args = parser.parse_args()
if args.subcommand == "test":
segmenter = DPSegmentLine(imweight=args.imweight,
bweight=args.bweight,
diagweight=args.diagweight,
r=args.r,
threshold=args.threshold,
debug=1)
ion()
gray()
for fname in args.files:
print fname
image = ocrolib.read_image_gray(fname)
segmentation = segmenter.charseg(image)
figure("output")
subplot(211)
imshow(image)
subplot(212)
morph.showlabels(segmentation)
raw_input()
else:
parser.print_help()
sys.exit(0)
def compute_geomaps(fnames,
shapedict,
old_model,
use_gt=1,
size=32,
debug=0,
old_order=1):
"""Given a shape dictionary and an existing line geometry
estimator, compute updated geometric maps for each entry
in the shape dictionary."""
if debug > 0:
gray()
ion()
shape = (shapedict.k, size, size)
bls = zeros(shape)
xls = zeros(shape)
count = 0
for fno, fname in enumerate(fnames):
if fno % 20 == 0: print fno, fname, count
if use_gt:
# don't use lines with many capital letters for training because
# they result in bad models
gt = ocrolib.read_text(ocrolib.fvariant(fname, "txt", "gt"))
if len(re.sub(r'[^A-Z]', '',
gt)) >= 0.3 * len(re.sub(r'[^a-z]', '', gt)):
continue
if len(re.sub(r'[^0-9]', '',
gt)) >= 0.3 * len(re.sub(r'[^a-z]', '', gt)):
continue
image = 1 - ocrolib.read_image_gray(fname)
if debug > 0 and fno % debug == 0:
clf()
subplot(411)
imshow(image)
try:
blp, xlp = old_model.lineFit(image, order=old_order)
except:
traceback.print_exc()
continue
blimage = zeros(image.shape)
h, w = image.shape
for x in range(w):
blimage[clip(int(polyval(blp, x)), 0, h - 1), x] = 1
xlimage = zeros(image.shape)
for x in range(w):
xlimage[clip(int(polyval(xlp, x)), 0, h - 1), x] = 1
if debug > 0 and fno % debug == 0:
subplot(413)
imshow(xlimage + 0.3 * image)
subplot(414)
imshow(blimage + 0.3 * image)
try:
seg = lineseg.ccslineseg(image)
except:
continue
if debug > 0 and fno % debug == 0:
subplot(412)
morph.showlabels(seg)
shape = None
for sub, transform, itransform_add in extract_chars(seg):
if shape is None: shape = sub.shape
assert sub.shape == shape
count += 1
best = shapedict.predict1(sub)
bls[best] += transform(blimage)
xls[best] += transform(xlimage)
if debug == 1: ginput(1, 100)
elif debug > 1: ginput(1, 0.01)
for i in range(len(bls)):
bls[i] *= bls[i].shape[1] * 1.0 / max(1e-6, sum(bls[i]))
for i in range(len(xls)):
xls[i] *= xls[i].shape[1] * 1.0 / max(1e-6, sum(xls[i]))
return bls, xls
