Repository navigation
Expand file tree
/
Copy pathfeature_extractor.py
More file actions
301 lines (249 loc) · 9.2 KB
/
Copy pathfeature_extractor.py
File metadata and controls
301 lines (249 loc) · 9.2 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
import cv2
import imutils
import numpy as np
from scipy import stats
from imutils import contours
from commonfunctions import *
from skimage.measure import regionprops
import glob
def returnToBGR(image):
return cv2.cvtColor(image,cv2.COLOR_GRAY2RGB)
def correct_skew(thresh):
coords = np.column_stack(np.where(thresh > 0))
angle = cv2.minAreaRect(coords)[-1]
if angle < -45:
angle = -(90 + angle)
else:
angle = -angle
(h, w) = thresh.shape[:2]
center = (w // 2, h // 2)
M = cv2.getRotationMatrix2D(center, angle, 1.0)
rotated = cv2.warpAffine(thresh, M, (w, h), flags=cv2.INTER_CUBIC, borderMode=cv2.BORDER_REPLICATE)
return rotated
def getVerticalProjection(img):
return (np.sum(img, axis = 0))//255
def getHorizontalProjection(img):
return (np.sum(img, axis = 1))//255
def DetectLines(line_img):
input_image = correct_skew(line_img)
_,thresh = cv2.threshold(input_image, 0, 255,cv2.THRESH_OTSU|cv2.THRESH_BINARY_INV)
rect_kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (25,5))
dilation = cv2.dilate(thresh, rect_kernel, iterations = 1)
contours_initial, hierarchy = cv2.findContours(dilation, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
im2 = input_image.copy()
Detected_Lines = []
# sort the contours
for method in ("right-to-left","top-to-bottom"):
contours_initial, boundingBoxes = contours.sort_contours(contours_initial, method=method)
for cnt in contours_initial:
x, y, w, h = cv2.boundingRect(cnt)
fx = x+w
fy = y+h
cv2.rectangle(im2, (x, y), (fx, fy), (100, 100, 100), 2)
trial_image = input_image[y:fy,x:fx]
Detected_Lines.append(trial_image)
return Detected_Lines
def DetectWords(line_img):
input_image = correct_skew(line_img)
_,thresh = cv2.threshold(input_image, 127, 255,cv2.THRESH_OTSU|cv2.THRESH_BINARY_INV)
rect_kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5))
dilation = cv2.dilate(thresh, rect_kernel, iterations = 1)
contours_initial, hierarchy = cv2.findContours(dilation, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
Detected_Words = []
# sort the contours
for method in ("right-to-left"):
contours_initial, boundingBoxes = contours.sort_contours(contours_initial, method=method)
for cnt in contours_initial:
x, y, w, h = cv2.boundingRect(cnt)
fx = x+w
fy = y+h
trial_image = input_image[y:fy,x:fx]
Detected_Words.append(trial_image)
return Detected_Words
def BaselineDetection(line_img): #4
PV = []
BLI = 0
_,thresh_img = cv2.threshold(line_img,127,255,cv2.THRESH_BINARY_INV)
thresh_img = np.asarray(thresh_img)
HP = getHorizontalProjection(thresh_img)
PV_Indices = (HP > np.roll(HP,1)) & (HP > np.roll(HP,-1))
for i in range(len(PV_Indices)):
if PV_Indices[i] == True:
PV.append(HP[i])
if len(PV) != 0:
MAX = max(PV)
else:
MAX = 36
for i in range(len(HP)):
if HP[i] == MAX:
BLI = i
#BLI = line_img.shape[0] - BLI
return BLI
####################################################################
def amrsFunction(word,BLI):
VP = getVerticalProjection(word)
copyWord = np.copy(word)
VP[VP < 2] = 0
if BLI < word.shape[0]:
copyWord[BLI, VP == 0] = 0
return copyWord
def extractLettersFromWord(word):
VP = getVerticalProjection(word)
VPIndices = np.where(VP == 0)[0]
temp = np.copy(VPIndices)
temp = np.diff(temp)
temp = np.insert(temp,0,0)
VPIndices = VPIndices[temp > 1]
return VPIndices
def showImagesFromSegments(word,listOfSegments):
i = 0
finalSegments = []
privious_final_segment_index = -1
for segment in listOfSegments:
if not (Is_excess_segment(word[:,i:segment])):
finalSegments.append(word[:,i:segment])
privious_final_segment_index += 1
i = segment
return finalSegments
def extractFromWord(word,BLI):
word = threshold(word)
word = word * 255
wordCopy = amrsFunction(word,BLI)
listOfSegmentations = extractLettersFromWord(wordCopy)
return showImagesFromSegments(word,listOfSegmentations)
def preprocessIntoWords(input_image):
if len(input_image.shape) == 3:
input_image = cv2.cvtColor(input_image, cv2.COLOR_BGR2GRAY)
input_image = correct_skew(input_image)
return DetectLines(input_image)
def Is_excess_segment (char):
HP = getHorizontalProjection (char)
HP[HP !=0] = 1
HPIndices = np.where(HP == 1)[0]
HP_Indices_diff = np.diff(HPIndices)
if (np.sum(char)//255) <= 4 and not( np.any(HP_Indices_diff > 1)):
return True
return False
def extractSeparateLettersWholeImage(input_image):
len_words = preprocessIntoWords(input_image)
all_words = []
for line in len_words:
BLI = BaselineDetection(line)
Detected_Words = DetectWords(line)
Detected_Words.reverse()
for word in Detected_Words:
SegmentedWord = extractFromWord(word,BLI)
all_words.append(SegmentedWord)
return all_words
#####################
#Abdelgawad
def Gabor_filter (img):
g_kernel = cv2.getGaborKernel((21, 21), 8.0, np.pi/4, 10.0, 0.5, 0, ktype=cv2.CV_32F)
filtered_img = cv2.filter2D(img, cv2.CV_8UC3, g_kernel)
filtered_img = np.array(filtered_img)
features = filtered_img.flatten()
return features
def horizontal_transitions(img):
horizontal_transition_count = 0
col_found_in = {}
for r in range(img.shape[0]):
for c in range(img.shape[1]-1):
if img[r,c] != img[r,c+1] and not( c in col_found_in.keys()) :
col_found_in[c] = 1
horizontal_transition_count+=1
return horizontal_transition_count
def vertcial_transisions (img):
vertical_transition_count = 0
row_found_in = {}
for c in range(img.shape[1]):
for r in range(img.shape[0]-1):
if img[r,c] != img[r+1,c] and not( r in row_found_in.keys()):
row_found_in[r] = 1
vertical_transition_count+=1
return vertical_transition_count
def number_of_transitions(img):
#img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
ht = horizontal_transitions(img)
vt = vertcial_transisions(img)
return ht,vt
#Mufeed
def Center_of_mass(letters):
properties = regionprops(letters, letters)
center_of_mass = properties[0].centroid
weighted_center_of_mass = properties[0].weighted_centroid
return (center_of_mass[1]/letters.shape[0]), (center_of_mass[0]/letters.shape[1])
def Black_ink_histogram(letter):# check needed
VP = getVerticalProjection(letter)
VP_ink = [VP[i*len(VP)//20] for i in range(20)]
HP = getHorizontalProjection(letter)
HP_ink = [HP[i*len(HP)//20] for i in range(20)]
return VP_ink, HP_ink
def Connected_Component(letter):
ret, labels, stats,center = cv2.connectedComponentsWithStats(letter)
return ret
#####################
#Salah
def ratio_of_colors(region1, region2, color1, color2):
'''
Gets ratios between 2 regions
color1 = 0, 1
color2 = 0, 1
0 --> black
1 --> white
'''
region1 = np.array(region1)
region2 = np.array(region2)
region1 = region1.flatten()
region2 = region2.flatten()
required1 = None
required2 = None
if color1 == 0:
required1 = region1[region1 == 0]
required1 = required1.shape[0]
else:
required1 = region1[region1 == 255]
required1 = required1.shape[0]
if color2 == 0:
required2 = region2[region2 == 0]
required2 = required2.shape[0]
else:
required2 = region2[region2 == 255]
required2 = required2.shape[0]
if required2 == 0:
return 0
return required1/required2
def height_over_width(image):
image = np.array(image)
return image.shape[0]/image.shape[1]
def divide_image_to_four_regions(image):
image = np.array(image)
img_cols = image.shape[1]
img_rows = image.shape[0]
mid_col = img_cols//2
mid_row = img_rows//2
first_region = image[0:mid_row,0:mid_col]
second_region = image[0:mid_row,mid_col:]
third_region = image[mid_row+1:,0:mid_col+1]
fourth_region = image[mid_row+1:,mid_col+1:]
return first_region, second_region, third_region, fourth_region
#####################
def ratiosBlackWhite(letter):
r1,r2,r3,r4 = divide_image_to_four_regions(letter)
f1 = ratio_of_colors(r1,r1,0,1)
f2 = ratio_of_colors(r2,r2,0,1)
f3 = ratio_of_colors(r3,r3,0,1)
f4 = ratio_of_colors(r4,r4,0,1)
f5 = ratio_of_colors(r1,r2,0,0)
f6 = ratio_of_colors(r3,r4,0,0)
f7 = ratio_of_colors(r1,r3,0,0)
f8 = ratio_of_colors(r2,r4,0,0)
f9 = ratio_of_colors(r1,r4,0,0)
f10 = ratio_of_colors(r2,r3,0,0)
return f1,f2,f3,f4,f5,f6,f7,f8,f9,f10
def count_holes(img, num_connected_parts): # count number of holes in each character
contours, hierarchy = cv2.findContours(img, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
if len(contours) == 0:
return 0
elif abs(len(contours) - (num_connected_parts-1)) < 0:
return 0
return abs(len(contours) - (num_connected_parts-1))