-
Notifications
You must be signed in to change notification settings - Fork 6
/
Global-Adaptive-Thresholding-Image-Segmentation.py
151 lines (151 loc) · 8.25 KB
/
Global-Adaptive-Thresholding-Image-Segmentation.py
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
import numpy as np
# OpenCV
import cv2 as cv
# To Read Folder of images
from os import listdir
# To Join the paths
from os.path import isfile, join
# To store DiceCoefficient of all images in Excel
import pandas as pd
# Global Adaptive Thresholding
def AdaptiveThresholding(Img, meanv, Epsilon):
output = np.zeros((rows, columns), dtype=np.uint8)
# infinte loop
while True:
# for two groups
i1 = []
i2 = []
# Dividing in two groups
for i in range(0, rows):
for j in range(0, columns):
if meanv > Img[i][j]:
i1.append(Img[i][j])
else:
i2.append(Img[i][j])
# finding mean of two groups separately
t1 = np.mean(i1)
t2 = np.mean(i2)
# Getting avg of two means
meanv2 = (t1+t2)/2
# subtraction new mean from old one
diff = meanv2 - meanv
# checking if diff is greater than epsilon then repeat the process for new value of mean
# otherwise new mean will be our threshold
Result = abs(diff)
if Result < Epsilon:
break
else:
meanv = meanv2
meanTH = meanv2
# Applying threshold
for i in range(0, rows):
for j in range(0, columns):
if Img[i][j] <= meanTH:
output[i][j] = 0
else:
output[i][j] = 255
return output
# Dice Coefficient to find the performance efficiency of Algorithm
def DiceCoefficient(output, actualImg):
TP = 0
FP = 0
FN = 0
TN = 0
for i in range(output.shape[0]):
for j in range(output.shape[1]):
if output[i][j] == 255 and actualImg[i][j] == 255:
TP = TP + 1
elif output[i][j] == 0 and actualImg[i][j] == 255:
FP = FP + 1
elif output[i][j] == 255 and actualImg[i][j] == 0:
FN = FN + 1
elif output[i][j] == 0 and actualImg[i][j] == 0:
TN = TN + 1
dice_coefficient = (2 * TP) / (FN + (2 * TP) + FP)
return dice_coefficient
# Function to remove the smallest objects using cca and only show the biggest one
def RemoveSmallRegions(output):
new_img = np.zeros_like(output)
for val in np.unique(output)[1:]:
mask = np.uint8(output == val) # step 3
labels, stats = cv.connectedComponentsWithStats(mask, 4)[1:3]
largest_label = 1 + np.argmax(stats[1:, cv.CC_STAT_AREA])
new_img[labels == largest_label] = val
return new_img
# DiceCoefficients for every image will append to this array
diceCoeff = []
# Images names in a specific folder are going to append in an array
mypath1 = r'C:\Users\ZAIN UL ABAIDIN\Desktop\NewRG'
onlyfiles1 = [f for f in listdir(mypath1) if isfile(join(mypath1,f))]
images1 = np.empty(len(onlyfiles1), dtype=object)
mypath2 = r'C:\Users\ZAIN UL ABAIDIN\Desktop\actualoutput'
onlyfiles2 = [f for f in listdir(mypath2) if isfile(join(mypath2, f))]
images2 = np.empty(len(onlyfiles2), dtype=object)
# To write every output in a folder with different name
strin = ['D002', 'D003', 'D004', 'D006', 'D008', 'D009', 'D010', 'D013', 'D014', 'D015',
'D016', 'D017', 'D018', 'D019', 'D020', 'D021', 'D022', 'D023', 'D024', 'D025',
'D027', 'D030', 'D031', 'D032', 'D033', 'D035', 'D036', 'D037', 'D038', 'D039',
'D040', 'D041', 'D042', 'D043', 'D044', 'D045', 'D047', 'D048', 'D049', 'D050',
'D057', 'D058', 'D061', 'D063', 'D064', 'D065', 'D075', 'D076', 'D078', 'D080',
'D085', 'D088', 'D090', 'D091', 'D092', 'D101', 'D103', 'D105', 'D107', 'D108',
'D112', 'D118', 'D120', 'D125', 'D126', 'D132', 'D133', 'D134', 'D135', 'D137',
'D138', 'D139', 'D140', 'D142', 'D143', 'D144', 'D146', 'D147', 'D149', 'D150',
'D152', 'D153', 'D154', 'D155', 'D156', 'D157', 'D159', 'D160', 'D161', 'D162',
'D164', 'D166', 'D168', 'D169', 'D170', 'D171', 'D173', 'D175', 'D176', 'D177',
'D182', 'D196', 'D197', 'D198', 'D199', 'D200', 'D203', 'D204', 'D206', 'D207',
'D208', 'D210', 'D211', 'D219', 'D226', 'D240', 'D242', 'D243', 'D251', 'D254',
'D256', 'D278', 'D279', 'D280', 'D284', 'D285', 'D304', 'D305', 'D306', 'D312',
'D328', 'D331', 'D339', 'D347', 'D348', 'D349', 'D356', 'D360', 'D364', 'D365',
'D367', 'D368', 'D369', 'D370', 'D371', 'D372', 'D374', 'D375', 'D378', 'D379',
'D380', 'D381', 'D382', 'D383', 'D384', 'D385', 'D386', 'D388', 'D389', 'D390',
'D392', 'D393', 'D394', 'D395', 'D396', 'D397', 'D398', 'D399', 'D400', 'D402',
'D403', 'D403', 'D404', 'D405', 'D406', 'D407', 'D408', 'D409', 'D410', 'D411',
'D413', 'D417', 'D418', 'D419', 'D420', 'D421', 'D423', 'D424', 'D425', 'D426',
'D427', 'D429', 'D430', 'D431', 'D432', 'D433', 'D434', 'D435', 'D436', 'D437']
# loop to read 200 images one by one
for i in range(0, len(onlyfiles1)):
# given image to extract lesion
orginalImg = cv.imread(join(mypath1, onlyfiles1[i]), 0)
# given Extracted lesion image
actualImg = cv.imread(join(mypath2, onlyfiles2[i]), 0)
# applying smoothing
orginalImg = cv.GaussianBlur(orginalImg, (15, 15), cv.BORDER_DEFAULT)
# applying closing to remove hair
kernel = np.ones((9, 9), np.uint8)
orginalImg = cv.morphologyEx(orginalImg, cv.MORPH_CLOSE, kernel, iterations=6)
rows = orginalImg.shape[0]
columns = orginalImg.shape[1]
meanv = np.absolute(np.mean(orginalImg))
Epsilon = 10
# Applying Global Adaptive Thresholding with Epsilon = 10
ATImg = AdaptiveThresholding(orginalImg, meanv, Epsilon)
ret, th1 = cv.threshold(ATImg, 80, 255, cv.THRESH_BINARY_INV)
# Removing Small Regions like corners
output = RemoveSmallRegions(th1)
df = DiceCoefficient(output, actualImg)
diceCoeff.append(df)
cv.imwrite(r'C:\Users\ZAIN UL ABAIDIN\Desktop\ATNew\IM' + strin[i] + '.bmp', output)
# Saving Dice Coeffcients to an Excel file
df1 = pd.DataFrame([diceCoeff]).transpose()
df1.columns = ['Dice Coefficient']
df1.index = ['img01', 'img02', 'img03', 'img04', 'img05', 'img06', 'img07', 'img08', 'img09', 'img10',
'img11', 'img12', 'img13', 'img14', 'img15', 'img16', 'img17', 'img18', 'img19', 'img20',
'img21', 'img22', 'img23', 'img24', 'img25', 'img26', 'img27', 'img28', 'img29', 'img30',
'img31', 'img32', 'img33', 'img34', 'img35', 'img36', 'img37', 'img38', 'img39', 'img40',
'img41', 'img42', 'img43', 'img44', 'img45', 'img46', 'img47', 'img48', 'img49', 'img50',
'img51', 'img52', 'img53', 'img54', 'img55', 'img56', 'img57', 'img58', 'img59', 'img60',
'img61', 'img62', 'img63', 'img64', 'img65', 'img66', 'img67', 'img68', 'img69', 'img70',
'img71', 'img72', 'img73', 'img74', 'img75', 'img76', 'img77', 'img78', 'img79', 'img80',
'img81', 'img82', 'img83', 'img84', 'img85', 'img86', 'img87', 'img88', 'img89', 'img90',
'img91', 'img92', 'img93', 'img94', 'img95', 'img96', 'img97', 'img98', 'img99', 'img100',
'img101', 'img102', 'img103', 'img104', 'img105', 'img106', 'img107', 'img108', 'img109', 'img110',
'img111', 'img112', 'img113', 'img114', 'img115', 'img116', 'img117', 'img118', 'img119', 'img120',
'img121', 'img122', 'img123', 'img124', 'img125', 'img126', 'img127', 'img128', 'img129', 'img130',
'img131', 'img132', 'img133', 'img134', 'img135', 'img136', 'img137', 'img138', 'img139', 'img140',
'img141', 'img142', 'img143', 'img144', 'img145', 'img146', 'img147', 'img148', 'img149', 'img150',
'img151', 'img152', 'img153', 'img154', 'img155', 'img156', 'img157', 'img158', 'img159', 'img160',
'img161', 'img162', 'img163', 'img164', 'img165', 'img166', 'img167', 'img168', 'img169', 'img170',
'img171', 'img172', 'img173', 'img174', 'img175', 'img176', 'img177', 'img178', 'img179', 'img180',
'img181', 'img182', 'img183', 'img184', 'img185', 'img186', 'img187', 'img188', 'img189', 'img190',
'img191', 'img192', 'img193', 'img194', 'img195', 'img196', 'img197', 'img198', 'img199', 'img200']
df1.to_excel('AdaptiveThresholdingDiceCoeff.xlsx')