-
Notifications
You must be signed in to change notification settings - Fork 7
/
Copy pathediting_classifier.py
296 lines (261 loc) · 12.3 KB
/
editing_classifier.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
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
from pydoc import doc
from templates import *
from templates_cls import *
from experiment_classifier import ClsModel
from torchvision.utils import *
import os
import glob
from tqdm import tqdm
from torch.utils.data import DataLoader
import sys
import dlib
import numpy as np
import skimage.io as io
import PIL
import PIL.Image
from PIL import Image
import scipy
from scipy.ndimage import gaussian_filter1d
import argparse
import imageio
def l2_norm(input, axis=1):
norm = torch.norm(input, 2, axis, True)
output = torch.div(input, norm)
return output
def tensor2pil(tensor: torch.Tensor) -> Image.Image:
x = tensor.squeeze(0).permute(1, 2, 0).add(1).mul(255).div(2).squeeze()
x = x.detach().cpu().numpy()
x = np.rint(x).clip(0, 255).astype(np.uint8)
return Image.fromarray(x)
def paste_image(coeffs, img, orig_image):
pasted_image = orig_image.copy().convert('RGBA')
projected = img.convert('RGBA').transform(orig_image.size, Image.PERSPECTIVE, coeffs, Image.BILINEAR)
pasted_image.paste(projected, (0, 0), mask=projected)
return pasted_image
def calc_alignment_coefficients(pa, pb):
matrix = []
for p1, p2 in zip(pa, pb):
matrix.append([p1[0], p1[1], 1, 0, 0, 0, -p2[0] * p1[0], -p2[0] * p1[1]])
matrix.append([0, 0, 0, p1[0], p1[1], 1, -p2[1] * p1[0], -p2[1] * p1[1]])
a = np.matrix(matrix, dtype=float)
b = np.array(pb).reshape(8)
res = np.dot(np.linalg.inv(a.T * a) * a.T, b)
return np.array(res).reshape(8)
def get_landmark(filepath, predictor, detector=None, fa=None):
"""get landmark with dlib
:return: np.array shape=(68, 2)
"""
if fa is not None:
image = io.imread(filepath)
lms, _, bboxes = fa.get_landmarks(image, return_bboxes=True)
if len(lms) == 0:
return None
return lms[0]
if detector is None:
detector = dlib.get_frontal_face_detector()
if isinstance(filepath, PIL.Image.Image):
img = np.array(filepath)
else:
img = dlib.load_rgb_image(filepath)
dets = detector(img)
for k, d in enumerate(dets):
shape = predictor(img, d)
break
else:
return None
t = list(shape.parts())
a = []
for tt in t:
a.append([tt.x, tt.y])
lm = np.array(a)
return lm
def compute_transform(filepath, predictor, detector=None, scale=1.0, fa=None):
lm = get_landmark(filepath, predictor, detector, fa)
if lm is None:
raise Exception(f'Did not detect any faces in image: {filepath}')
lm_chin = lm[0: 17] # left-right
lm_eyebrow_left = lm[17: 22] # left-right
lm_eyebrow_right = lm[22: 27] # left-right
lm_nose = lm[27: 31] # top-down
lm_nostrils = lm[31: 36] # top-down
lm_eye_left = lm[36: 42] # left-clockwise
lm_eye_right = lm[42: 48] # left-clockwise
lm_mouth_outer = lm[48: 60] # left-clockwise
lm_mouth_inner = lm[60: 68] # left-clockwise
# Calculate auxiliary vectors.
eye_left = np.mean(lm_eye_left, axis=0)
eye_right = np.mean(lm_eye_right, axis=0)
eye_avg = (eye_left + eye_right) * 0.5
eye_to_eye = eye_right - eye_left
mouth_left = lm_mouth_outer[0]
mouth_right = lm_mouth_outer[6]
mouth_avg = (mouth_left + mouth_right) * 0.5
eye_to_mouth = mouth_avg - eye_avg
# Choose oriented crop rectangle.
x = eye_to_eye - np.flipud(eye_to_mouth) * [-1, 1]
x /= np.hypot(*x)
x *= max(np.hypot(*eye_to_eye) * 2.0, np.hypot(*eye_to_mouth) * 1.8)
x *= scale
y = np.flipud(x) * [-1, 1]
c = eye_avg + eye_to_mouth * 0.1
return c, x, y
def crop_image(filepath, output_size, quad, enable_padding=False):
x = (quad[3] - quad[1]) / 2
qsize = np.hypot(*x) * 2
# read image
if isinstance(filepath, PIL.Image.Image):
img = filepath
else:
img = PIL.Image.open(filepath)
transform_size = output_size
# Shrink.
shrink = int(np.floor(qsize / output_size * 0.5))
if shrink > 1:
rsize = (int(np.rint(float(img.size[0]) / shrink)), int(np.rint(float(img.size[1]) / shrink)))
img = img.resize(rsize, PIL.Image.ANTIALIAS)
quad /= shrink
qsize /= shrink
# Crop.
border = max(int(np.rint(qsize * 0.1)), 3)
crop = (int(np.floor(min(quad[:, 0]))), int(np.floor(min(quad[:, 1]))), int(np.ceil(max(quad[:, 0]))),
int(np.ceil(max(quad[:, 1]))))
crop = (max(crop[0] - border, 0), max(crop[1] - border, 0), min(crop[2] + border, img.size[0]),
min(crop[3] + border, img.size[1]))
if (crop[2] - crop[0] < img.size[0] or crop[3] - crop[1] < img.size[1]):
img = img.crop(crop)
quad -= crop[0:2]
# Pad.
pad = (int(np.floor(min(quad[:, 0]))), int(np.floor(min(quad[:, 1]))), int(np.ceil(max(quad[:, 0]))),
int(np.ceil(max(quad[:, 1]))))
pad = (max(-pad[0] + border, 0), max(-pad[1] + border, 0), max(pad[2] - img.size[0] + border, 0),
max(pad[3] - img.size[1] + border, 0))
if enable_padding and max(pad) > border - 4:
pad = np.maximum(pad, int(np.rint(qsize * 0.3)))
img = np.pad(np.float32(img), ((pad[1], pad[3]), (pad[0], pad[2]), (0, 0)), 'reflect')
h, w, _ = img.shape
y, x, _ = np.ogrid[:h, :w, :1]
mask = np.maximum(1.0 - np.minimum(np.float32(x) / pad[0], np.float32(w - 1 - x) / pad[2]),
1.0 - np.minimum(np.float32(y) / pad[1], np.float32(h - 1 - y) / pad[3]))
blur = qsize * 0.02
img += (scipy.ndimage.gaussian_filter(img, [blur, blur, 0]) - img) * np.clip(mask * 3.0 + 1.0, 0.0, 1.0)
img += (np.median(img, axis=(0, 1)) - img) * np.clip(mask, 0.0, 1.0)
img = PIL.Image.fromarray(np.uint8(np.clip(np.rint(img), 0, 255)), 'RGB')
quad += pad[:2]
# Transform.
img = img.transform((transform_size, transform_size), PIL.Image.QUAD, (quad + 0.5).flatten(), PIL.Image.BILINEAR)
if output_size < transform_size:
img = img.resize((output_size, output_size), PIL.Image.ANTIALIAS)
return img
parser = argparse.ArgumentParser()
parser.add_argument("--batch_size", type=int, default=4, help="batch size for editing process")
parser.add_argument("--T", type=int, default=1000, help="total diffusion step")
parser.add_argument("--max_num", type=int, default=None, help="maximum number of frames to edit, 1~max_num")
parser.add_argument("--attribute", type=str, default="Mustache", required=True, help="editing attribute in CelebAttrDataset")
parser.add_argument("--scale", type=float, default=0.3, required=True, help="editing magnitude, + and - both possible")
parser.add_argument("--video_path", type=str, default="sample_video", required=True, help="path to video frames")
parser.add_argument("--normalize", default=True, action='store_true', help="normalize the feature space during editing")
args = parser.parse_args()
gpus = [0]
device = 'cuda:0'
cls_conf = diffusion_video_autoencoder_cls(gpus)
conf = diffusion_video_autoencoder(gpus)
state = torch.load(f'checkpoints/{conf.name}/epoch=51-step=1000000.ckpt', map_location='cpu')
model = LitModel(conf)
model.load_state_dict(state['state_dict'], strict=False)
model.ema_model.eval()
model.ema_model.to(device);
cls_model = ClsModel(cls_conf)
state = torch.load(f'checkpoints/{cls_conf.name}/last.ckpt',
map_location='cpu')
print('latent step:', state['global_step'])
cls_model.load_state_dict(state['state_dict'], strict=False);
cls_model.to(device);
print(CelebAttrDataset.id_to_cls)
log_dir = f'editing_classifier/{args.video_path}'
if not os.path.exists(log_dir):
os.mkdir(log_dir)
if not os.path.exists(os.path.join(log_dir, "crop")):
os.mkdir(os.path.join(log_dir, "crop"))
if not os.path.exists(os.path.join(log_dir, "recon")):
os.mkdir(os.path.join(log_dir, "recon"))
predictor = dlib.shape_predictor("pretrained_models/shape_predictor_68_face_landmarks.dat")
detector = dlib.get_frontal_face_detector()
with torch.no_grad():
images = []
dir_name = args.video_path
assert os.path.isdir(dir_name), '%s is not a valid directory' % dir_name
for fname in sorted(os.listdir(dir_name)):
path = os.path.join(dir_name, fname)
fname = fname.split('.')[0]
images.append((fname, path))
cs, xs, ys = [], [], []
for _, path in images:
c, x, y = compute_transform(path, predictor, detector=detector, scale=1.0)
cs.append(c)
xs.append(x)
ys.append(y)
cs = np.stack(cs)
xs = np.stack(xs)
ys = np.stack(ys)
cs = gaussian_filter1d(cs, sigma=1.0, axis=0)
xs = gaussian_filter1d(xs, sigma=3.0, axis=0)
ys = gaussian_filter1d(ys, sigma=3.0, axis=0)
quads = np.stack([cs - xs - ys, cs - xs + ys, cs + xs + ys, cs + xs - ys], axis=1)
quads = list(quads)
orig_images = []
for quad, (_, path) in tqdm(zip(quads, images), total=len(quads)):
crop = crop_image(path, 1024, quad.copy())
crop.save(f'{log_dir}/crop/%s.jpg' % path.split('/')[-1].split('.')[0])#, quality=100, subsampling=0)
orig_image = Image.open(path)
orig_images.append(orig_image)
image_size = 256
inverse_transforms = [
calc_alignment_coefficients(quad + 0.5, [[0, 0], [0, image_size], [image_size, image_size], [image_size, 0]])
for quad in quads]
data = ImageDataset(f'{log_dir}/crop', image_size=conf.img_size, exts=['jpg', 'JPG', 'png'], do_augment=False, sort_names=True, max_num=args.max_num)
dataloader = DataLoader(data, batch_size=args.batch_size, shuffle=False, drop_last=False)
conds = []
for batch in dataloader:
imgs = batch['img']
indices = batch['index']
cond = model.ema_model.encoder.id_forward(imgs.to(device))
conds.append(cond)
cond = torch.cat(conds, dim=0)
avg_cond = torch.mean(cond, dim=0, keepdim=True) #.expand(len(x_start), -1)
video_frames = []
if args.normalize:
avg_cond_norm = cls_model.normalize(avg_cond)
for batch in dataloader:
imgs = batch['img']
indices = batch['index']
avg_latent = model.ema_model.encoder.forward_with_id(avg_cond.expand(len(imgs), -1), imgs.to(device))
avg_xT = model.encode_stochastic(imgs.to(device), avg_latent, T=args.T)
mask = model.ema_model.encoder.face_mask(imgs.to(device), for_video=True)
avg_img_recon = model.render(avg_xT, avg_latent, T=args.T)
ori = (imgs + 1) / 2
for index in range(len(imgs)):
file_name = data.paths[indices[index]]
save_image(ori[index], f'{log_dir}/recon/orig_{file_name}')
save_image(avg_xT[index], f'{log_dir}/recon/avg_xT_{file_name}')
save_image(avg_img_recon[index], f'{log_dir}/recon/avg_recon_{file_name}')
cls_id = CelebAttrDataset.cls_to_id[args.attribute]
if args.normalize:
avg_cond2 = avg_cond_norm + args.scale * math.sqrt(512) * F.normalize(cls_model.classifier.weight[cls_id][None, :], dim=1)
avg_cond2 = l2_norm(cls_model.denormalize(avg_cond2))
else:
avg_cond2 = l2_norm(avg_cond + args.scale * math.sqrt(512) * cls_model.classifier.weight[cls_id][None, :])
if not os.path.exists(f'{log_dir}/{args.attribute}_{args.scale:.2f}'):
os.mkdir(f'{log_dir}/{args.attribute}_{args.scale:.2f}')
avg_latent2 = model.ema_model.encoder.forward_with_id(avg_cond2.expand(len(imgs), -1), imgs.to(device))
avg_img = model.render(avg_xT, avg_latent2, T=args.T)
for index in range(len(imgs)):
file_name = data.paths[indices[index]]
# save_image(avg_img[index], f'{log_dir}/{args.attribute}_{args.scale:.2f}/avg_mani_{file_name}')
paste_bg = avg_img[index].unsqueeze(0) * mask[index].unsqueeze(0) + ((imgs[index].to(device).unsqueeze(0) + 1) / 2 * (1 - mask[index].unsqueeze(0)))
save_image(paste_bg[0], f'{log_dir}/{args.attribute}_{args.scale:.2f}/paste_avg_mani_{file_name}')
paste_bg_crop = tensor2pil((paste_bg[0] * 2) - 1)
paste_bg_pasted_image = paste_image(inverse_transforms[indices[index]], paste_bg_crop, orig_images[indices[index]])
paste_bg_pasted_image = paste_bg_pasted_image.convert('RGB')
video_frames.append(paste_bg_pasted_image)
paste_bg_pasted_image.save(f'{log_dir}/{args.attribute}_{args.scale:.2f}/paste_final_avg_mani_{file_name}')
imageio.mimwrite(f'{log_dir}/{args.attribute}_{args.scale:.2f}/out.mp4', video_frames, fps=20, output_params=['-vf', 'fps=20'])