-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathtest.py
211 lines (173 loc) · 6.4 KB
/
test.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
205
206
import os
import argparse
from tqdm import tqdm
import cv2
# import ffmpeg
import numpy as np
import torch
from torch.utils.data import DataLoader
from configs import cfg
from metrics import psnr, ssim_metric
from utils.data_utils import select_dataset
import lpips
import time
from validate import load_render, mkdir
loss_fn_alex = lpips.LPIPS(net="alex").cuda() # best forward scores
loss_fn_vgg = lpips.LPIPS(
net="vgg"
).cuda() # closer to "traditional" perceptual loss, when used for optimization
loss_fn_alex.eval()
loss_fn_vgg.eval()
def myinfer(infer_dataset, render, save_dir, epoch=0):
render.eval()
psnr_wMask_list = []
psnr_woMask_list = []
ssim_list = []
lpips_alex_list = []
lpips_vgg_list = []
img_dir = f"{save_dir}/{epoch}/img"
rendering_dir = f"{save_dir}/{epoch}/rendering"
gt_dir = f"{save_dir}/{epoch}/ground_truth"
acc_dir = f"{save_dir}/{epoch}/acc"
depth_dir = f"{save_dir}/{epoch}/depth"
mkdir(img_dir)
mkdir(acc_dir)
mkdir(depth_dir)
mkdir(gt_dir)
mkdir(rendering_dir)
# with torch.no_grad():
for batch_idx, batch in enumerate(tqdm(infer_dataset)):
# batch['frame'][...] = 0
real_frame = batch["frame"][0]
## fix frame code
# batch["frame"][...] = 59
if "save_name" in batch:
save_name = batch["save_name"][0]
else:
frame_index = batch["frame_index"].item()
view_index = batch["cam_ind"].item()
save_name = "frame{:04d}_view{:04d}".format(frame_index, view_index)
# import pdb;pdb.set_trace()
results = render.render_view(batch)
color_img_0 = results["coarse_color"]
color_img_0 = torch.clamp(color_img_0, min=0.0, max=1.0)
depth_img_0 = results["coarse_depth"]
acc_map_0 = results["coarse_acc"]
color_gt = batch["img"][0]
H, W = color_gt.shape[:2]
mask_at_box = batch["mask_at_box"][0].bool().reshape(H, W)
psnr_wMask = psnr(color_img_0, color_gt, mask_at_box)
psnr_woMask = psnr(color_img_0, color_gt)
ssim_ = ssim_metric(
color_img_0.cpu().numpy(), color_gt.cpu().numpy(), mask_at_box
)
pred = (
(2 * color_img_0 - 1).permute(2, 0, 1)[None].float().flip(1)
) ### TO RGB ,(-1,1)
gt = (2 * color_gt - 1).permute(2, 0, 1)[None].float().flip(1)
pred = pred.cuda()
gt = gt.cuda()
with torch.no_grad():
lpips_alex = loss_fn_alex(pred, gt).squeeze().cpu()
lpips_vgg = loss_fn_vgg(pred, gt).squeeze().cpu()
psnr_wMask_list.append(psnr_wMask)
psnr_woMask_list.append(psnr_woMask)
ssim_list.append(ssim_)
lpips_alex_list.append(lpips_alex)
lpips_vgg_list.append(lpips_vgg)
img_path = os.path.join(img_dir, f"{save_name}.png")
rendering_path = os.path.join(rendering_dir, f"{save_name}.png")
gt_path = os.path.join(gt_dir, f"{save_name}.png")
acc_path = os.path.join(acc_dir, f"{save_name}.png")
depth_path = os.path.join(depth_dir, f"{save_name}.png")
rendering = color_img_0.numpy() * 255
gt = batch["img"].squeeze().numpy() * 255
cat_img = np.concatenate((rendering, gt), axis=1)
cv2.imwrite(img_path, cat_img)
cv2.imwrite(rendering_path, rendering)
cv2.imwrite(gt_path, gt)
depth_img_0 = np.repeat(depth_img_0.numpy(), 3, axis=2) * 255
cv2.imwrite(depth_path, depth_img_0)
acc_map_0 = np.repeat(acc_map_0.numpy(), 3, axis=2) * 255
cv2.imwrite(acc_path, acc_map_0)
psnr_wMask_mean = np.array(psnr_wMask_list).mean()
psnr_woMask_mean = np.array(psnr_woMask_list).mean()
ssim_mean = np.array(ssim_list).mean()
lpips_alex_mean = np.array(lpips_alex_list).mean()
lpips_vgg_mean = np.array(lpips_vgg_list).mean()
print("epoch", epoch)
print("psnr_wMask_mean", psnr_wMask_mean)
print("psnr_woMask_mean", psnr_woMask_mean)
print("ssim_mean", ssim_mean)
print("lpips_alex_mean", lpips_alex_mean)
print("lpips_vgg_mean", lpips_vgg_mean)
return {
"psnr_wMask": psnr_wMask_mean,
"psnr_woMask": psnr_woMask_mean,
"ssim": ssim_mean,
"lpips_alex": lpips_alex_mean,
"lpips_vgg": lpips_vgg_mean,
}
def save_img(imgs, img_dir):
if not os.path.exists(img_dir):
os.makedirs(img_dir)
for idx in range(len(imgs)):
img_path = os.path.join(img_dir, "%06d.jpg" % idx)
cv2.imwrite(img_path, imgs[idx])
def img2vid(img_dir, output_path):
(
ffmpeg.input("%s/*.jpg" % img_dir, pattern_type="glob", framerate=15)
.output(output_path)
.run()
)
if __name__ == "__main__":
save_root = "./TEST"
parser = argparse.ArgumentParser(description="infer")
parser.add_argument(
"-c",
"--config",
default="",
help="set the config file path to train the network",
)
parser.add_argument("--exp", type=str, default="test")
parser.add_argument("--ckpt", type=str, required=True)
args = parser.parse_args()
epoch = int(os.path.basename(args.ckpt).split(".")[0].split("_")[-1])
save_dir = os.path.join(save_root, args.exp)
# Load config
training_config = args.config
assert os.path.exists(training_config), "training config does not exist."
cfg.merge_from_file(training_config)
novel_view_dataset, novel_pose_dataset = select_dataset(cfg, formal_test=True)
novel_view_dataloader = DataLoader(
novel_view_dataset, batch_size=1, shuffle=False, num_workers=4
)
novel_pose_dataloader = DataLoader(
novel_pose_dataset, batch_size=1, shuffle=False, num_workers=4
)
render = load_render(
args.ckpt, cfg, canonical_vertex=novel_view_dataset.canonical_vertex
)
# import pdb;pdb.set_trace()
# print(cfg)
print("novel view:")
out1 = myinfer(
novel_view_dataloader,
render,
save_dir=os.path.join(save_dir, "novel_view"),
epoch=epoch,
)
print("novel pose:")
render.net.set_light_center(
torch.tensor(cfg.TEST.light_center)
)
render.net.nerf.w = 0
out2 = myinfer(
novel_pose_dataloader,
render,
save_dir=os.path.join(save_dir, "novel_pose"),
epoch=epoch,
)
# import pdb;pdb.set_trace()
# save_img(out, os.path.join(save_dir, "imgs"))
# img2vid(save_dir, os.path.join(save_dir, f"{args.exp}.mp4"))