-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathStatistic.py
198 lines (162 loc) · 6.65 KB
/
Statistic.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
import scipy.io as scio
import os
import pandas as pd
import json
import numpy as np
import sys
from shutil import copyfile
class Logger(object):
def __init__(self, filename='default.log', stream=sys.stdout):
self.terminal = stream
self.log = open(filename, 'w')
def write(self, message):
self.terminal.write(message)
self.log.write(message)
def flush(self):
pass
def mkdir(path):
folder = os.path.exists(path)
if not folder: #判断是否存在文件夹如果不存在则创建为文件夹
os.makedirs(path) #makedirs 创建文件时如果路径不存在会创建这个路径
def MaxMinNormalization(x):
"""[0,1] normaliaztion"""
x = (x - np.min(x)) / (np.max(x) - np.min(x))
return x
Model="SINet_V2_convnext_L2_edgetask_fusion"
# Model="SINet_V2_convnext_small"
DataPath = './Dataset/TestDataset/'
Datasets = ['CAMO', 'CHAMELEON','COD10K']
mkdir(f"./Statistics/{Model}/")
sys.stdout = Logger(f"./Statistics/{Model}/{Model}_statistic.log", sys.stdout)
print("************************************************************")
dataFile = f'./EvaluationResults/{Model}/COD10K-mat/{Model}_img.mat'
data = scio.loadmat(dataFile)
caminfo=pd.read_excel(io='/media/perry/E/DataSets/COD10K-v2/COD10K-v2/Info/Statistics-CAM.xlsx')
caminfo=json.loads(caminfo.to_json(orient="index",force_ascii=False))
for i in range(len(caminfo)):
ImgName=caminfo[str(i)]["ImgName"]
caminfo[ImgName] = caminfo.pop(str(i))
del caminfo[ImgName]["ImgName"]
gtPath = DataPath + Datasets[2] + '/GT/'
imgFiles=[]
imgFiles_temp = os.listdir(gtPath)
for filename in imgFiles_temp:
if os.path.splitext(filename)[1] == '.png':
imgFiles.append(filename)
#建立字典库
img_dict=[]
category_dict=[]
subcategory_dict=[]
#添加指标和标识符等特征
for i,imgfile in enumerate(imgFiles):
category=imgfile.split("-")[3]
if category not in category_dict:
category_dict.append(category)
subcategory=imgfile.split("-")[5]
if subcategory not in subcategory_dict:
subcategory_dict.append(subcategory)
CAMinfo=caminfo[imgfile[:-4]]
CAMkey=list(CAMinfo.keys())
CAM=np.zeros(7)
for j in range(len(CAMinfo)):
if CAMinfo[CAMkey[j]] == None:
CAM[j]=0
else:
CAM[j]=1
Smeasure=data["Smeasure"][0,i]
wFmeasure=data["wFmeasure"][0,i]
MAE=data["MAE"][0,i]
Mean_Emeasure=np.mean(data["threshold_Emeasure"],1)[i]
img_dict.append({"imgfile":imgfile,"category":category,"subcategory":subcategory,"CAM":CAM,\
"Smeasure":Smeasure,"wFmeasure":wFmeasure,"MAE":MAE,"Mean_Emeasure":Mean_Emeasure})
category_dict.sort()
subcategory_dict.sort()
for img_item in img_dict:
category=np.zeros(5)
subcategory=np.zeros(69)
cat_ind=category_dict.index(img_item["category"])
scat_ind=subcategory_dict.index(img_item["subcategory"])
category[cat_ind]=1
subcategory[scat_ind]=1
img_item["category"]=category
img_item["subcategory"]=subcategory
#多元线性回归
df = pd.DataFrame(img_dict)
x=[np.array(df["category"].tolist()),np.array(df["subcategory"].tolist()),np.array(df["CAM"].tolist())]
y=[np.array(df["Smeasure"].tolist()),np.array(df["Mean_Emeasure"].tolist()),np.array(df["wFmeasure"].tolist()),np.array(df["MAE"].tolist())]
x_lable=["category","subcategory","CAM"]
y_lable=["Smeasure","Mean_Emeasure","wFmeasure","MAE"]
# Res=[]
for i in range(len(x)):
# res=[]
for j in range(len(y)):
# 标准化
y[j]=MaxMinNormalization(y[j])
data_x=pd.DataFrame(x[i])
data_y=pd.Series(y[j])
# 查看是否符合高斯分布 画散点图和直方图(非高斯分布不适用z-score标准化)
# import matplotlib.pyplot as plt
# plt.switch_backend('agg')
# fig = plt.figure(figsize = (10,6))
# ax1 = fig.add_subplot(2,1,1) # 创建子图1
# ax1.scatter(data_y.index, data_y.values)
# plt.grid()
# ax2 = fig.add_subplot(2,1,2) # 创建子图2
# data_y.hist(bins=30,alpha = 0.5,ax = ax2)
# data_y.plot(kind = 'kde', secondary_y=True,ax = ax2)
# plt.grid()
# plt.savefig(f"./Statistics/{Model}/{Model}_"+y_lable[j]+"_"+ str(j) +".jpg")
# res.append(np.linalg.lstsq(data_x,data_y,rcond=None))
# Res.append(res)
theta = np.dot(np.dot(np.linalg.inv(np.dot(data_x.T, data_x)), data_x.T), data_y)
pred_y = np.dot(theta, data_x.T)
print('x_lable',x_lable[i],'y_lable',y_lable[j])
print('权重',theta) # 权重
loss = np.mean(np.array(data_y-pred_y)) # 残差
print('残差',loss)
print('')
#四种指标排序并进行基础统计
x_lable=["category","subcategory","CAM"]
y_lable=["Smeasure","Mean_Emeasure","wFmeasure","MAE"]
# 对于四种指标
for i in range(len(y_lable)):
df=df.sort_values(by=[y_lable[i]])
stastic_data=[df["imgfile"].tolist(),np.array(df[y_lable[i]].tolist()),np.array(df["category"].tolist()),np.array(df["subcategory"].tolist()),np.array(df["CAM"].tolist())]
category_num=np.zeros(5)
subcategory_num=np.zeros(69)
cam_num=np.zeros(7)
errfile_list=[]
if y_lable[i]!="MAE": #MAE越小越好，取最大10%统计错误样本特征
# 逐个累加标签信息
for k in range(df.shape[0]):
if k < np.round(df.shape[0]*0.1):
errfile_list.append([stastic_data[0][k],k])
category_num +=stastic_data[2][k]
subcategory_num +=stastic_data[3][k]
cam_num +=stastic_data[4][k]
else:
break
else:
for k in range(df.shape[0]):
if k > np.round(df.shape[0]*0.9):
errfile_list.append([stastic_data[0][k],k])
category_num +=stastic_data[2][k]
subcategory_num +=stastic_data[3][k]
cam_num +=stastic_data[4][k]
print('y_lable',y_lable[i])
print('category_num',category_num)
print('subcategory_num',subcategory_num)
print('cam_num',cam_num)
print('')
# 写入错误图片
for camind,camtype in enumerate(CAMkey):
errorimg_path=f"./Statistics/errorimg/{Model}/{Model}_10par_{y_lable[i]}/"+str(camind)+"_"+camtype+"/"
mkdir(errorimg_path)
for errfile,k in errfile_list:
if stastic_data[4][k][camind]==1:
predimgpath=f"res/{Model}/COD10K/"+errfile
copyfile(predimgpath, errorimg_path+errfile)
print("************************************************************")
# img_dict.sort(key=lambda k: k["A_MAE"])
# for i in sorted (img_dict) :
# print ((i, key_value[i]), end =" ")