-
Notifications
You must be signed in to change notification settings - Fork 5
/
evaluation.py
204 lines (175 loc) · 6.75 KB
/
evaluation.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
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
import argparse
import sys
from PIL import Image
import numpy as np
import torch
import math
parser = argparse.ArgumentParser()
parser.add_argument('--gtdir', type=str,
default='/home/zhangcb/Desktop/VOCtrainval/VOCdevkit/VOC2012/SegmentationObjectContour_gt',
help='groundTruthDir')
parser.add_argument('--resultdir', type=str,
default='/home/zhangcb/Desktop/resnet50/result/04m_09d_21h/predict',
help='resultDir')
parser.add_argument('--txt', type=str,
default='/home/zhangcb/Desktop/resnet50/data/val.txt',
help='val.txt')
args, _ = parser.parse_known_args(sys.argv[1:])
file_root = args.txt
base_root_result = args.resultdir
base_root_mask = args.gtdir
def compute_fusion_matrix(match_result, gt, threshold):
result = torch.tensor(np.array(match_result)).to('cuda:0')
gt = torch.tensor(np.array(gt)).to('cuda:0')
result[result < threshold] = 0
result[result >= threshold] = 1
gt[gt != 255] = 0
gt[gt == 255] = 1
tmp1 = result - gt
tmp2 = result + gt
TP = (tmp2 == 2).sum()
TN = (tmp2 == 0).sum()
FP = (tmp1 == 1).sum()
FN = (tmp1 == -1).sum()
return TP.item(), FP.item(), TN.item(), FN.item()
def main():
precision = np.zeros(256)
recall = np.zeros(256)
result_path = []
mask_path = []
with open(file_root, 'r') as f:
for line in f:
pic_root_img = base_root_result + '/' + line.split('\n')[0] + '.jpg'
result_path.append(pic_root_img)
pic_root_mask = base_root_mask + '/' + line.split('\n')[0] + '.png'
mask_path.append(pic_root_mask)
for i in range(len(mask_path)):
#result_path[i] = base_root_result + '/' + '2007_000129.jpg'
#mask_path[i] = base_root_mask + '/' + '2007_000129.png'
result = np.array(Image.open(result_path[i]))
mask = np.array(Image.open(mask_path[i]))
#result = NMS(result)
for threshold in range(256):
TP, FP, TN, FN = compute_fusion_matrix(result, mask, threshold)
if TP + FP == 0:
p = 0
else:
p = TP / (TP + FP)
if TP + FN == 0:
r = 0
else:
r = TP / (TP + FN)
precision[threshold] += p
recall[threshold] += r
precision /= len(mask_path)
recall /= len(mask_path)
return precision, recall
# 计算梯度幅值
def gradients(new_gray):
"""
:type: image which after smooth
:rtype:
dx: gradient in the x direction
dy: gradient in the y direction
M: gradient magnitude
theta: gradient direction
"""
W, H = new_gray.shape
dx = np.zeros([W - 1, H - 1])
dy = np.zeros([W - 1, H - 1])
M = np.zeros([W - 1, H - 1])
theta = np.zeros([W - 1, H - 1])
for i in range(W - 1):
for j in range(H - 1):
dx[i, j] = new_gray[i + 1, j] - new_gray[i, j]
dy[i, j] = new_gray[i, j + 1] - new_gray[i, j]
# 图像梯度幅值作为图像强度值
M[i, j] = np.sqrt(np.square(dx[i, j]) + np.square(dy[i, j]))
# 计算 θ - artan(dx/dy)
theta[i, j] = math.atan(dx[i, j] / (dy[i, j] + 0.000000001))
return dx, dy, M, theta
def NMS(new_gary):
dx, dy, M, theta = gradients(new_gary)
d = np.copy(M)
W, H = M.shape
NMS = np.copy(d)
NMS[0, :] = NMS[W - 1, :] = NMS[:, 0] = NMS[:, H - 1] = 0
for i in range(1, W - 1):
for j in range(1, H - 1):
# 如果当前梯度为0,该点就不是边缘点
if M[i, j] == 0:
NMS[i, j] = 0
else:
gradX = dx[i, j] # 当前点 x 方向导数
gradY = dy[i, j] # 当前点 y 方向导数
gradTemp = d[i, j] # 当前梯度点
# 如果 y 方向梯度值比较大,说明导数方向趋向于 y 分量
if np.abs(gradY) > np.abs(gradX):
weight = np.abs(gradX) / np.abs(gradY) # 权重
grad2 = d[i - 1, j]
grad4 = d[i + 1, j]
# 如果 x, y 方向导数符号一致
# 像素点位置关系
# g1 g2
# c
# g4 g3
if gradX * gradY > 0:
grad1 = d[i - 1, j - 1]
grad3 = d[i + 1, j + 1]
# 如果 x,y 方向导数符号相反
# 像素点位置关系
# g2 g1
# c
# g3 g4
else:
grad1 = d[i - 1, j + 1]
grad3 = d[i + 1, j - 1]
# 如果 x 方向梯度值比较大
else:
weight = np.abs(gradY) / np.abs(gradX)
grad2 = d[i, j - 1]
grad4 = d[i, j + 1]
# 如果 x, y 方向导数符号一致
# 像素点位置关系
# g3
# g2 c g4
# g1
if gradX * gradY > 0:
grad1 = d[i + 1, j - 1]
grad3 = d[i - 1, j + 1]
# 如果 x,y 方向导数符号相反
# 像素点位置关系
# g1
# g2 c g4
# g3
else:
grad1 = d[i - 1, j - 1]
grad3 = d[i + 1, j + 1]
# 利用 grad1-grad4 对梯度进行插值
gradTemp1 = weight * grad1 + (1 - weight) * grad2
gradTemp2 = weight * grad3 + (1 - weight) * grad4
# 当前像素的梯度是局部的最大值,可能是边缘点
if gradTemp >= gradTemp1 and gradTemp >= gradTemp2:
NMS[i, j] = gradTemp
else:
# 不可能是边缘点
NMS[i, j] = 0
return NMS
if __name__ == '__main__':
precision, recall = main()
precision, recall = np.array(precision), np.array(recall)
np.save('/home/zhangcb/Desktop/resnet50/util/precision.npy', precision)
np.save('/home/zhangcb/Desktop/resnet50/util/recall.npy', recall)
Fscore = np.true_divide(2 * np.multiply(precision, recall), precision + recall)
arg_threshold = np.argmax(Fscore)
Fscore = Fscore.max()
print('max F-score = ', Fscore, 'threshold = ', arg_threshold)
import matplotlib.pyplot as plt
plt.figure()
plt.ylim(0, 1.1)
plt.xlim(0, 1.1)
plt.ylabel('precision')
plt.xlabel('recall')
plt.plot(recall, precision)
plt.show()
#print(precision, recall)