/
synergynet_util.py
193 lines (157 loc) · 6.11 KB
/
synergynet_util.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
"""SynergyNet utils"""
import random
import pickle
import mindspore
from mindspore import ops
from mindspore import Tensor
import os
import numpy as np
import argparse
class Crop:
"""
Input:
Tensor: shape(3, 120, 120),
Return:
Tensor: shape(3, 120, 120)
"""
def __init__(self, maximum, std=None, prob=0.01, mode='test'):
self.maximum = maximum
self.std = std
self.prob = prob
self.type_li = [1, 2, 3, 4, 5, 6, 7]
self.switcher = {
1: self.lup,
2: self.rup,
3: self.ldown,
4: self.rdown,
5: self.lhalf,
6: self.rhalf,
7: self.center
}
self.mode = mode
self.zero = ops.Zeros()
def get_params(self, img):
h = img.shape[1]
w = img.shape[2]
crop_margins = self.maximum
rand = random.random()
return crop_margins, h, w, rand
def lup(self, img, h, w):
new_img = self.zero((3, h, w), mindspore.float32)
new_img[:, :h // 2, :w // 2] = img[:, :h // 2, :w // 2]
return new_img
def rup(self, img, h, w):
new_img = self.zero((3, h, w), mindspore.float32)
new_img[:, :h // 2, w // 2:] = img[:, :h // 2, w // 2:]
return new_img
def ldown(self, img, h, w):
new_img = self.zero((3, h, w), mindspore.float32)
new_img[:, h // 2:, :w // 2] = img[:, h // 2:, :w // 2]
return new_img
def rdown(self, img, h, w):
new_img = self.zero((3, h, w), mindspore.float32)
new_img[:, :h // 2, :w // 2] = img[:, :h // 2, :w // 2]
return new_img
def lhalf(self, img, h, w):
new_img = self.zero((3, h, w), mindspore.float32)
new_img[:, :, :w // 2] = img[:, :, :w // 2]
return new_img
def rhalf(self, img, h, w):
new_img = self.zero((3, h, w), mindspore.float32)
new_img[:, :, w // 2:] = img[:, :, w // 2:]
return new_img
def center(self, img, h, w):
new_img = self.zero((3, h, w), mindspore.float32)
new_img[:, h // 4: -h // 4, w // 4: -w // 4] = img[:, h // 4: -h // 4, w // 4: -w // 4]
return new_img
def __call__(self, img, gt=None):
img_tensor = Tensor(img, dtype=mindspore.float32)
crop_margins, h, w, rand = self.get_params(img_tensor)
crop_backgnd = self.zero((3, h, w), mindspore.float32)
crop_backgnd[:, crop_margins:h - 1 * crop_margins, crop_margins:w - 1 * crop_margins] = \
img_tensor[:, crop_margins: h - crop_margins, crop_margins: w - crop_margins]
# random center crop
if (rand < self.prob) and (self.mode == 'train'):
func = self.switcher.get(random.randint(1, 7))
crop_backgnd = func(crop_backgnd, h, w)
# center crop
if self.mode == 'test':
crop_backgnd[:, crop_margins:h - 1 * crop_margins, crop_margins:w - 1 * crop_margins] = \
img_tensor[:, crop_margins: h - crop_margins, crop_margins: w - crop_margins]
crop_backgnd = crop_backgnd.asnumpy()
return crop_backgnd
def make_abs_path(d):
return os.path.join(os.path.dirname(os.path.realpath(__file__)), d)
def _get_suffix(filename):
"""a.jpg -> jpg"""
pos = filename.rfind('.')
if pos == -1:
return ''
return filename[pos + 1:]
def _load(fp):
suffix = _get_suffix(fp)
if suffix == 'npy':
return np.load(fp)
elif suffix == 'pkl':
return pickle.load(open(fp, 'rb')) # 和 dump() 函数相对应，用于将二进制对象文件转换成 Python 对象
def mkdir(d):
"""only works on *nix system"""
if not os.path.isdir(d) and not os.path.exists(d):
os.system('mkdir -p {}'.format(d))
class Compose_GT(object):
def __init__(self, transforms):
self.transforms = transforms
def __call__(self, img, gt):
for t in self.transforms:
img = t(img)
img = Tensor.from_numpy(img)
return img, gt
def str2bool(v):
if v.lower() in ('yes', 'true', 't', 'y', '1'):
return True
elif v.lower() in ('no', 'false', 'f', 'n', '0'):
return False
else:
raise argparse.ArgumentTypeError('Boolean value expected')
class AverageMeter(object):
def __init__(self):
self.reset()
def reset(self):
self.val = 0
self.avg = 0
self.sum = 0
self.count = 0
def update(self, val, n=1):
self.val = val
self.sum += val * n
self.count += n
self.avg = self.sum / self.count
class ParamsPack:
"""Parameter package"""
def __init__(self):
try:
d = make_abs_path('E:/LAB/threeD/SynergyNet-main/3dmm_data')
self.keypoints = _load(os.path.join(d, 'keypoints_sim.npy'))
# PCA basis for shape, expression, texture
self.w_shp = _load(os.path.join(d, 'w_shp_sim.npy'))
self.w_exp = _load(os.path.join(d, 'w_exp_sim.npy'))
# param_mean and param_std are used for re-whitening
meta = _load(os.path.join(d, 'param_whitening.pkl'))
self.param_mean = meta.get('param_mean')
self.param_std = meta.get('param_std')
# mean values
self.u_shp = _load(os.path.join(d, 'u_shp.npy'))
self.u_exp = _load(os.path.join(d, 'u_exp.npy'))
self.u = self.u_shp + self.u_exp
self.w = np.concatenate((self.w_shp, self.w_exp), axis=1)
# base vector for landmarks
self.w_base = self.w[self.keypoints]
self.w_norm = np.linalg.norm(self.w, axis=0)
self.w_base_norm = np.linalg.norm(self.w_base, axis=0)
self.u_base = self.u[self.keypoints].reshape(-1, 1)
self.w_shp_base = self.w_shp[self.keypoints]
self.w_exp_base = self.w_exp[self.keypoints]
self.std_size = 120
self.dim = self.w_shp.shape[0] // 3
except:
raise RuntimeError('Missing data')