-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathindoor_pred.py
193 lines (167 loc) · 9.66 KB
/
indoor_pred.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
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import itertools
import cv2
from torchvision.models import mobilenet_v2
from sklearn.metrics import confusion_matrix
from sklearn import metrics
import random
from PIL import Image
import argparse
import numpy as np
import torch
from torchvision import transforms, models
from sklearn.model_selection import train_test_split
import torch.nn as nn
import os.path
import matplotlib.pyplot as plt
import os
parser = argparse.ArgumentParser(description='PyTorch Training Script')
parser.add_argument('--batch_size', type=int, default=64, metavar='N',
help='input batch size for training (default: 64)')
parser.add_argument('--test_batch', type=int, default=64, metavar='N',
help='input batch size for testing (default: 64)')
parser.add_argument('--epochs', type=int, default=150, metavar='N',
help='number of epochs to train (default: 10)')
parser.add_argument('--no-cuda', action='store_true', default=False,
help='disables CUDA training')
parser.add_argument('--seed', type=int, default=1, metavar='S',
help='random seed (default: 1)')
parser.add_argument('--log-interval', type=int, default=25, metavar='N',
help='how many batches to wait before logging training status')
parser.add_argument('--resume', type=bool, default=True, metavar='N',
help='resume from the last weights')
torch.cuda.empty_cache()
args = parser.parse_args()
args.cuda = not args.no_cuda and torch.cuda.is_available()
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
NumCell = 30 # number of cells
NumClass = 6 # number of classes except background class
polygonList = [[0, 0, 151, 0, 52, 54, 0, 54], [151, 0, 254, 0, 191, 54, 52, 54], [254, 0, 359, 0, 334, 54, 191, 54],
[359, 0, 424, 0, 424, 54, 334, 54], [489, 0, 424, 0, 424, 54, 514, 54],
[594, 0, 489, 0, 514, 54, 657, 54], [697, 0, 594, 0, 657, 54, 796, 54],
[848, 0, 697, 0, 796, 54, 848, 54], [0, 54, 52, 54, 0, 82], [52, 54, 191, 54, 133, 102, 0, 102, 0, 82],
[191, 54, 334, 54, 312, 102, 133, 102], [334, 54, 424, 54, 424, 102, 312, 102],
[514, 54, 424, 54, 424, 102, 536, 102], [657, 54, 514, 54, 536, 102, 715, 102],
[796, 54, 657, 54, 715, 102, 848, 102, 848, 82], [848, 54, 796, 54, 848, 82], [0, 102, 133, 102, 0, 212],
[133, 102, 312, 102, 277, 178, 266, 179, 255, 180, 244, 181, 234, 182, 223, 183, 213, 184, 203, 185, 193,
186, 183, 187, 173, 188, 163, 189, 153, 190, 145, 191, 137, 192, 129, 193, 121, 194, 113, 195, 106, 196,
99, 197, 93, 198, 86, 199, 80, 200, 74, 201, 68, 202, 62, 203, 57, 204, 51, 205, 46, 206, 41, 207, 36,
208, 31, 209, 27, 210, 22, 211, 17, 212, 13, 213, 8, 214, 3, 215, 0, 216, 0, 212],
[312, 102, 424, 102, 424, 178, 277, 178], [536, 102, 424, 102, 424, 178, 571, 178],
[715, 102, 536, 102, 571, 178, 582, 179, 593, 180, 604, 181, 614, 182, 625, 183, 635, 184, 645, 185, 655,
186, 665, 187, 675, 188, 685, 189, 695, 190, 703, 191, 711, 192, 719, 193, 727, 194, 735, 195, 742, 196,
749, 197, 755, 198, 762, 199, 768, 200, 774, 201, 780, 202, 786, 203, 791, 204, 797, 205, 802, 206, 807,
207, 812, 208, 817, 209, 821, 210, 826, 211, 831, 212, 835, 213, 840, 214, 845, 215, 848, 216, 848,
212], [848, 102, 715, 102, 848, 212],
[277, 178, 266, 179, 255, 180, 244, 181, 234, 182, 223, 183, 213, 184, 203, 185, 193, 186, 183, 187, 173,
188, 163, 189, 153, 190, 145, 191, 137, 192, 129, 193, 121, 194, 113, 195, 106, 196, 99, 197, 93, 198,
86, 199, 80, 200, 74, 201, 68, 202, 62, 203, 57, 204, 51, 205, 46, 206, 41, 207, 36, 208, 31, 209, 27,
210, 22, 211, 17, 212, 13, 213, 8, 214, 3, 215, 0, 216, 0, 338, 203, 338],
[277, 178, 424, 178, 424, 338, 203, 338], [571, 178, 424, 178, 424, 338, 645, 338],
[571, 178, 582, 179, 593, 180, 604, 181, 614, 182, 625, 183, 635, 184, 645, 185, 655, 186, 665, 187, 675,
188, 685, 189, 695, 190, 703, 191, 711, 192, 719, 193, 727, 194, 735, 195, 742, 196, 749, 197, 755, 198,
762, 199, 768, 200, 774, 201, 780, 202, 786, 203, 791, 204, 797, 205, 802, 206, 807, 207, 812, 208, 817,
209, 821, 210, 826, 211, 831, 212, 835, 213, 840, 214, 845, 215, 848, 216, 848, 338, 645, 338],
[0, 338, 203, 338, 137, 480, 0, 480], [203, 338, 424, 338, 424, 480, 137, 480],
[645, 338, 424, 338, 424, 480, 711, 480], [848, 338, 645, 338, 711, 480, 848, 480]]
model = mobilenet_v2() # load the model
model.classifier[1] = nn.Linear(1280, NumCell * (NumClass + 1))
model.load_state_dict(torch.load("./weights/mobilenet_indoor.pth.tar"))
save_name = 'mobilenet_indoor' # name of the model
class_names = ["Sofa", "Wall", "Pillar", "People", "Door", "Others", "Road", "Background"]
binary_class_names = ["Risk", "Road"]
val_test_trans = transforms.Compose(([
transforms.Resize((224, 224)),
transforms.ToTensor(), # divides by 255
]))
class MultiLabelRGBataSet(torch.utils.data.Dataset):
def __init__(self, imgspath, imgslist, annotationpath):
self.imgslist = imgslist
self.imgspath = imgspath
self.annotationpath = annotationpath
def __len__(self):
return len(self.imgslist)
def __getitem__(self, index):
ipath = os.path.join(self.imgspath, self.imgslist[index])
img = cv2.imread(ipath)
(filename, extension) = os.path.splitext(ipath)
filename = os.path.basename(filename)
annotation = os.path.join(self.annotationpath, filename + ".txt")
label = np.loadtxt(annotation, dtype=np.int64)
label = torch.tensor(label, dtype=torch.float32)
return img, label, filename
img_dir = 'images'
label_dir = 'labels'
img_list = os.listdir(img_dir)
train_img, Val_Test = train_test_split(img_list, test_size=0.3, random_state=2)
val_img, test_img = train_test_split(Val_Test, test_size=0.6666, random_state=2)
testSet = MultiLabelRGBataSet(img_dir, test_img, label_dir)
def polygon_center(points):
x = [p[0] for p in points]
y = [p[1] for p in points]
length = len(points)
area = 0.5 * sum(x[i-1]*y[i] - x[i]*y[i-1] for i in range(length))
center_x = sum((x[i-1] + x[i]) * (x[i-1]*y[i] - x[i]*y[i-1]) for i in range(length)) / (6*area)
center_y = sum((y[i-1] + y[i]) * (x[i-1]*y[i] - x[i]*y[i-1]) for i in range(length)) / (6*area)
return int(center_x), int(center_y)
def pred_plot(frame, original, output):
orig = original.detach().numpy()
output = output.detach().numpy()
pred = np.where(output > 0.5, 1, 0).tolist()
cell = 0
normal = np.asarray([0] * NumClass + [1])
for polygon in polygonList:
points = [(x, y) for x, y in zip(polygon[0::2], polygon[1::2])]
# each = pred[cell:cell + NumClass + 1]
eachScore = output[cell:cell + NumClass + 1]
each = orig[cell:cell + NumClass + 1]
if not (each == normal).all():
index = [i for i, x in enumerate(each) if x == 1]
if len(index) == 0:
index.append(eachScore.argmax())
if eachScore.argmax() == NumClass:
continue
center_x, center_y = polygon_center(points)
poly_area = cv2.contourArea(np.array(points, np.int32))
default_text_scale = 0.4 # 这是默认的字体大小，可以根据你的需要进行调整
texts = [] # 用于存储所有的文本和对应的大小
max_text_len = len(index) # 计算最长的文本长度
if max_text_len > 1:
text_scale = default_text_scale * min(1, np.sqrt(poly_area) / max_text_len)
else:
text_scale = min(max(poly_area / 10000, 0.3), 0.6)
for i in index:
text_size, _ = cv2.getTextSize(class_names[i], cv2.FONT_HERSHEY_SIMPLEX, text_scale, 2)
texts.append((class_names[i], text_size, text_scale))
text_origin = [center_x, center_y - sum(text[1][1] for text in texts) // 2]
line_spacing = 0.7 # 行间距，可以根据需要调整
for text, text_size, text_scale in texts:
text_origin[0] = center_x - text_size[0] // 2 # 每行的x坐标需要重新计算以保证居中
text_origin[1] += int(text_size[1] * line_spacing) # y坐标加上当前行文本的高度的一部分
cv2.putText(frame, text, tuple(text_origin), cv2.FONT_HERSHEY_SIMPLEX, text_scale, (255, 255, 255), 1)
text_origin[1] += int(text_size[1] * line_spacing) # y坐标再加上当前行文本的高度的一部分，为下一行文本做准备
cv2.polylines(frame, [np.array(points, np.int32)], isClosed=True, color=(0, 255, 255), thickness=3)
cell += NumClass + 1
# cv2.imshow("frame", frame)
# cv2.waitKey(0)
return frame
preprocess = transforms.Compose([transforms.ToTensor()])
def test():
model.eval()
with torch.no_grad():
for batch_idx, (image, target, filename) in enumerate(testSet):
resize_image = cv2.resize(image, (224, 224))
input_tensor = preprocess(resize_image)
input_batch = input_tensor.unsqueeze(0)
output = model(input_batch)
output = torch.sigmoid(output)
frame = pred_plot(image, torch.Tensor.cpu(target), output[0])
cv2.imwrite(os.path.join(path, filename + ".jpg"), frame)
current_path = os.getcwd()
path = os.path.join(current_path, save_name)
if not os.path.exists(path):
os.makedirs(path)
test()