-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_ck_bp4d_pretrain.py
146 lines (125 loc) · 6.87 KB
/
test_ck_bp4d_pretrain.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
import argparse
import os
import torch.utils.data as util_data
from tqdm import tqdm
from models.jaav2_base.jaanet import JaaNetv2
from utils.util import *
from data_ck.data_list import ImageList
from data import pre_process as prep
def main(config):
use_gpu = torch.cuda.is_available()
config.use_gpu = use_gpu
# prepare data
dsets = {}
dset_loaders = {}
dsets['test'] = ImageList(crop_size=config.crop_size, path=config.test_path_prefix, phase='test',
transform=prep.image_test(crop_size=config.crop_size),
target_transform=prep.land_transform(img_size=config.crop_size,
flip_reflect=np.loadtxt(
config.flip_reflect))
)
dset_loaders['test'] = util_data.DataLoader(dsets['test'], batch_size=config.eval_batch_size,
shuffle=False, num_workers=config.num_workers)
# set network modules
net = JaaNetv2(config)
if use_gpu:
net = net.cuda()
net.eval()
# save result
if not os.path.exists(config.res_path_prefix):
os.mkdir(config.res_path_prefix)
if not os.path.exists(config.res_path_prefix + config.model_prefix):
os.mkdir(config.res_path_prefix + config.model_prefix)
res_file = open(config.res_path_prefix + config.model_prefix + '/result.txt', 'w')
au_pred_file = config.res_path_prefix + config.model_prefix + '/au_pred.txt'
loader = dset_loaders['test']
for i, batch in enumerate(tqdm(loader)):
input, land, biocular, au = batch
if use_gpu:
input, land, au = input.cuda(), land.cuda(), au.cuda()
aus_output, local_aus_output, align_output = net(input)
if i == 0:
all_local_output = local_aus_output.data.cpu().float()
all_output = aus_output.data.cpu().float()
all_au = au.data.cpu().float()
all_pred_land = align_output.data.cpu().float()
all_land = land.data.cpu().float()
else:
all_local_output = torch.cat((all_local_output, local_aus_output.data.cpu().float()), 0)
all_output = torch.cat((all_output, aus_output.data.cpu().float()), 0)
all_au = torch.cat((all_au, au.data.cpu().float()), 0)
all_pred_land = torch.cat((all_pred_land, align_output.data.cpu().float()), 0)
all_land = torch.cat((all_land, land.data.cpu().float()), 0)
AUoccur_pred_prob = all_output.data.numpy()
local_AUoccur_pred_prob = all_local_output.data.numpy()
AUoccur_actual = all_au.data.numpy()
pred_land = all_pred_land.data.numpy()
GT_land = all_land.data.numpy()
# save AUoccur_pred_prob
np.savetxt(au_pred_file, AUoccur_pred_prob, fmt='%f', delimiter='\t')
local_f1score_arr, local_acc_arr, f1score_arr, acc_arr, mean_error, failure_rate = au_detection_eval_ck_bp4d(
AUoccur_pred_prob, local_AUoccur_pred_prob, AUoccur_actual, pred_land, GT_land
)
#print('=============================')
#print(local_f1score_arr.shape)
#print(local_acc_arr.shape)
#print(f1score_arr.shape)
#print(acc_arr.shape)
#print(mean_error.shape)
#print(failure_rate.shape)
# au_detection_eval_ck_bp4d
# bp4d: 1,2,4,6,7,10,12,14,15,17,23,24
# ck+ : 1,2,-,6,7,--,12,14,15,17,23,24 # del pred of [:, 2] and [:, 5]
# record result
line1 = "Test model, train on {}, test on data: {}".format(config.model_prefix, config.test_path_prefix)
line2 = "F1 score of each au is: " \
"au1={}, au2={}, au6={}, au7={}, au12={}, " \
"au15={}, au17={}, au23={}, au24={}".format(
f1score_arr[0], f1score_arr[1], f1score_arr[2], f1score_arr[3], f1score_arr[4],
f1score_arr[5], f1score_arr[6], f1score_arr[7], f1score_arr[8])
line3 = "Avarage F1 score is: avg={}".format(f1score_arr.mean())
line4 = "Local F1 score of each au is: " \
"au1={}, au2={}, au6={}, au7={}, au12={}, " \
"au15={}, au17={}, au23={}, au24={}".format(
local_f1score_arr[0], local_f1score_arr[1], local_f1score_arr[2], local_f1score_arr[3], local_f1score_arr[4],
local_f1score_arr[5], local_f1score_arr[6], local_f1score_arr[7], local_f1score_arr[8])
line5 = "Local Avarage F1 score is: avg={}".format(local_f1score_arr.mean())
line6 = "Landmark mean error is: mean_error= {}".format(mean_error)
line7 = "Acc of each au is: " \
"au1={}, au2={}, au6={}, au7={}, au12={}, " \
"au15={}, au17={}, au23={}, au24={}".format(
acc_arr[0], acc_arr[1], acc_arr[2], acc_arr[3], acc_arr[4],
acc_arr[5], acc_arr[6], acc_arr[7], acc_arr[8])
line8 = "Local acc is: local_acc_arr={}, Acc is: acc_arr={}".format(local_acc_arr.mean(),
acc_arr.mean())
for line in [line1, line2, line3, line4, line5, line6, line7, line8]:
print(line+'\n')
res_file.write(line+'\n')
res_file.close()
if __name__ == "__main__":
parser = argparse.ArgumentParser()
# Misc
parser.add_argument('--training', type=bool, default=False, help='training or testing')
parser.add_argument('--use_gpu', type=bool, default=True, help='default use gpu')
parser.add_argument('--gpu_id', type=str, default='0', help='device id to run')
parser.add_argument('--eval_batch_size', type=int, default=8, help='mini-batch size for evaluation')
parser.add_argument('--num_workers', type=int, default=8)
# Model configuration.
parser.add_argument('--crop_size', type=int, default=176, help='crop size for images')
parser.add_argument('--map_size', type=int, default=44, help='size for attention maps')
parser.add_argument('--au_num', type=int, default=12, help='number of AUs')
parser.add_argument('--land_num', type=int, default=49, help='number of landmarks')
parser.add_argument('--unit_dim', type=int, default=8, help='unit dims')
parser.add_argument('--lambda_au', type=float, default=1, help='weight for AU detection loss')
parser.add_argument('--lambda_land', type=float, default=0.5, help='weight for landmark detection loss')
# Directories.
parser.add_argument('--model_prefix', type=str, default='CK_BP4D_combine_2_3')
parser.add_argument('--pretrain_prefix', type=str, default='./models/weights/BP4D_combine_2_3')
parser.add_argument('--train_path_prefix', type=str, default='./data_ck/lists/CK')
parser.add_argument('--test_path_prefix', type=str, default='./data_ck/lists/CK')
parser.add_argument('--flip_reflect', type=str, default='data/list/reflect_49.txt')
parser.add_argument('--res_path_prefix', type=str, default='results/CK_BP4D_combine2_3/')
config = parser.parse_args()
os.environ['CUDA_VISIBLE_DEVICES'] = config.gpu_id
print(config)
main(config)