def make_compare_file():
f=open('./corrected_docs/Samp_'+str(tno)+'/compare_list_new.txt','w')
g=open('./corrected_docs/Samp_'+str(tno)+'/output_file_new.txt','w')
# img=cv2.imread('./Example/dc_books_page.png',0)
path='./corrected_docs/Samp_'+str(tno)+'/*.png'
url=glob.glob(path)
img=cv2.imread(url[0],0)
# img=cv2.imread('./Samp_3/samp3.png',0)
if(img==None):
print 'image does\'nt exist'
exit()
img = pp.preprocess(img)
# im=img
# im,rot = pp.skew_correction(img)
line = pp.find_lines(img.copy())
# print len(linene)
label_list=train.label_unicode()
i=0
num=[]
for l in line:
for w in l.word_list:
for c in w.char_list:
# num.append((str(i),label_list[int(c.label)]))
tup=label_list[int(c.label)]
f.write(tup+'\n')
g.write(tup)
# cv2.imwrite('./Samp_22/samp/'+str(i)+'.png',c.data)
i+=1
g.write(' ')
g.write('\n')
f.close()
g.close()
def make_modified_file():
f=open('./compare_list.txt','r')
g=open('./output_file.txt','w')
img=cv2.imread('./Example/dc_books_page.png',0)
if(img==None):
print url+' does\'nt exist'
exit()
img = pp.preprocess(img)
im,rot = pp.skew_correction(img)
line = pp.find_lines(im.copy())
# print len(linene)
label_list=train.label_unicode()
q=f.readlines()
i=0
num=[]
for l in line:
for w in l.word_list:
for c in w.char_list:
# num.append((str(i),label_list[int(c.label)]))
tup=label_list[int(c.label)]
if(q[i][:-1]!=tup):
print tup
# f.write(tup+'\n')
g.write(tup)
# cv2.imwrite('samp/'+str(i)+'.png',c.data)
i+=1
g.write(' ')
g.write('\n')
f.close()
g.close()
def make_modified_file():
f = open('./compare_list.txt', 'r')
g = open('./output_file.txt', 'w')
img = cv2.imread('./Example/dc_books_page.png', 0)
if (img == None):
print url + ' does\'nt exist'
exit()
img = pp.preprocess(img)
im, rot = pp.skew_correction(img)
line = pp.find_lines(im.copy())
# print len(linene)
label_list = train.label_unicode()
q = f.readlines()
i = 0
num = []
for l in line:
for w in l.word_list:
for c in w.char_list:
# num.append((str(i),label_list[int(c.label)]))
tup = label_list[int(c.label)]
if (q[i][:-1] != tup):
print tup
# f.write(tup+'\n')
g.write(tup)
# cv2.imwrite('samp/'+str(i)+'.png',c.data)
i += 1
g.write(' ')
g.write('\n')
f.close()
g.close()
def make_compare_file():
f = open('./corrected_docs/Samp_' + str(tno) + '/compare_list_new.txt',
'w')
g = open('./corrected_docs/Samp_' + str(tno) + '/output_file_new.txt', 'w')
# img=cv2.imread('./Example/dc_books_page.png',0)
path = './corrected_docs/Samp_' + str(tno) + '/*.png'
url = glob.glob(path)
img = cv2.imread(url[0], 0)
# img=cv2.imread('./Samp_3/samp3.png',0)
if (img == None):
print 'image does\'nt exist'
exit()
img = pp.preprocess(img)
# im=img
# im,rot = pp.skew_correction(img)
line = pp.find_lines(img.copy())
# print len(linene)
label_list = train.label_unicode()
i = 0
num = []
for l in line:
for w in l.word_list:
for c in w.char_list:
# num.append((str(i),label_list[int(c.label)]))
tup = label_list[int(c.label)]
f.write(tup + '\n')
g.write(tup)
# cv2.imwrite('./Samp_22/samp/'+str(i)+'.png',c.data)
i += 1
g.write(' ')
g.write('\n')
f.close()
g.close()
def recognize_block(im):
line = find_lines(im)
# print len(linene)
label_list = train.label_unicode()
i = 0
string = ''
#selecting each line
for l in line:
cv2.imwrite('temp/zline_' + str(i) + '.png', l.data)
string = string + '\n'
j = 0
#selecting words in a line
for w in l.word_list:
#cv2.imwrite('zword_'+str(i)+'_word_'+str(j)+'.png',w.data)
string = string + ' '
j += 1
k = 0
c = 0
#Formatting characters in the word
while (c < len(w.char_list)):
char = w.char_list[c]
try:
#checking whether the input is ' or " or ,
if (label_list[int(char.label)] in ['\'', ',']):
char2 = w.char_list[c + 1]
if (label_list[int(char2.label)] in ['\'', ',']):
string = string + '\"'
c += 1
else:
string = string + label_list[int(char.label)]
#checking whether the input is ൈ or െ
elif (label_list[int(char.label)] in ['െ', 'േ', '്ര']):
char2 = w.char_list[c + 1]
if (label_list[int(char2.label)] in ['െ', '്ര']):
char3 = w.char_list[c + 2]
string = string + label_list[int(char3.label)]
c += 1
string = string + label_list[int(char2.label)]
string = string + label_list[int(char.label)]
c += 1
else:
string = string + label_list[int(char.label)]
except IndexError:
string = string + label_list[int(char.label)]
# cv2.imwrite('output/zcline_'+str(i)+'_word_'+str(j)+'_c_'+str(k)+str(int(w.char_list[c].label))+'.png',w.char_list[c].data)
k += 1
c += 1
i += 1
return string
def recognize_block(im):
line = find_lines(im)
# print len(linene)
label_list=train.label_unicode()
i=0
string=''
#selecting each line
for l in line:
cv2.imwrite('temp/zline_'+str(i)+'.png',l.data)
string=string+'\n'
j=0
#selecting words in a line
for w in l.word_list:
#cv2.imwrite('zword_'+str(i)+'_word_'+str(j)+'.png',w.data)
string=string+' '
j+=1
k=0
c=0
#Formatting characters in the word
while(c<len(w.char_list)):
char= w.char_list[c]
try:
#checking whether the input is ' or " or ,
if(label_list[int(char.label)]in ['\'',',']):
char2=w.char_list[c+1]
if(label_list[int(char2.label)]in ['\'',',']):
string=string+'\"'
c+=1
else:
string=string+label_list[int(char.label)]
#checking whether the input is ൈ or െ
elif(label_list[int(char.label)]in ['െ','േ','്ര']):
char2=w.char_list[c+1]
if(label_list[int(char2.label)]in ['െ','്ര']):
char3=w.char_list[c+2]
string=string+label_list[int(char3.label)]
c+=1
string=string+label_list[int(char2.label)]
string=string+label_list[int(char.label)]
c+=1
else:
string=string+label_list[int(char.label)]
except IndexError:
string=string+label_list[int(char.label)]
# cv2.imwrite('output/zcline_'+str(i)+'_word_'+str(j)+'_c_'+str(k)+str(int(w.char_list[c].label))+'.png',w.char_list[c].data)
k+=1
c+=1
i+=1
return string
def recognize_block(im):
line = pp.find_lines(im)
# print len(linene)
label_list=train.label_unicode()
i=0
string='word:'
for l in line:
# cv2.imwrite('zline_'+str(i)+'.png',l.data)
string=string+'\n'
j=0
for w in l.word_list:
# cv2.imwrite('zword_'+str(i)+'_word_'+str(j)+'.png',w.data)
string=string+' '
j+=1
k=0
c=0
while(c<len(w.char_list)):
char= w.char_list[c]
try:
if(label_list[int(char.label)]in ['\'',',1',',2']):
char2=w.char_list[c+1]
if(label_list[int(char2.label)]in ['\'',',1',',2']):
string=string+'\"'
c+=1
else:
string=string+label_list[int(char.label)]
elif(label_list[int(char.label)]in [',1',',2']):
string=string+','
elif(label_list[int(char.label)]in ['ൾ2','ൾ']):
string=string+'ൾ'
elif(label_list[int(char.label)]in ['െ','േ','്ര']):
char2=w.char_list[c+1]
if(label_list[int(char2.label)]in ['െ','്ര']):
char3=w.char_list[c+2]
string=string+label_list[int(char3.label)]
c+=1
string=string+label_list[int(char2.label)]
string=string+label_list[int(char.label)]
c+=1
else:
string=string+label_list[int(char.label)]
except IndexError:
string=string+label_list[int(char.label)]
# cv2.imwrite('output/zcline_'+str(i)+'_word_'+str(j)+'_c_'+str(k)+str(int(w.char_list[c].label))+'.png',w.char_list[c].data)
k+=1
c+=1
i+=1
return string
def recognize_block(im):
line = pp.find_lines(im)
# print len(linene)
label_list = train.label_unicode()
i = 0
string = ''
for l in line:
# cv2.imwrite('zline_'+str(i)+'.png',l.data)
# string=string+'\n'
j = 0
for w in l.word_list:
# cv2.imwrite('zword_'+str(i)+'_word_'+str(j)+'.png',w.data)
string = string + ' '
j += 1
k = 0
c = 0
while (c < len(w.char_list)):
char = w.char_list[c]
try:
if (label_list[int(char.label)] in ['\'', ',']):
char2 = w.char_list[c + 1]
if (label_list[int(char2.label)] in ['\'', ',']):
string = string + '\"'
c += 1
else:
string = string + label_list[int(char.label)]
elif (label_list[int(char.label)] in ['െ', 'േ', '്ര']):
char2 = w.char_list[c + 1]
if (label_list[int(char2.label)] in ['െ', '്ര']):
char3 = w.char_list[c + 2]
string = string + label_list[int(char3.label)]
c += 1
string = string + label_list[int(char2.label)]
string = string + label_list[int(char.label)]
c += 1
else:
string = string + label_list[int(char.label)]
except IndexError:
string = string + label_list[int(char.label)]
cv2.imwrite(
'output/zcline_' + str(i) + '_word_' + str(j) + '_c_' +
str(k) + str(int(w.char_list[c].label)) + '.png',
w.char_list[c].data)
k += 1
c += 1
i += 1
return string