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
|
#!/usr/bin/env python2.7
import os
from subprocess import call
import sys
import getopt
import numpy
import argparse
import torch
import PIL
import PIL.Image
from datetime import datetime
from network import Network, process
from moviepy.video.io.ffmpeg_reader import FFMPEG_VideoReader
from moviepy.video.io.ffmpeg_writer import FFMPEG_VideoWriter
##########################################################
FPS = 25
parser = argparse.ArgumentParser()
parser.add_argument('--model', type=str, default='lf') # l1 or lf
parser.add_argument('--first', type=str, default='./images/first.png')
parser.add_argument('--second', type=str, default='./images/second.png')
parser.add_argument('--out', type=str, default='./result.png')
parser.add_argument('--video', action='store_true')
parser.add_argument('--video-out', type=str, default=datetime.now().strftime("sepconv_%Y%m%d_%H%M.mp4"))
parser.add_argument('--steps', type=int, default=0)
parser.add_argument('--dilate', type=int, default=1)
parser.add_argument('--smooth', action='store_true')
parser.add_argument('--mix-videos', action='store_true')
parser.add_argument('--average-videos', action='store_true')
parser.add_argument('--a-offset', type=int, default=0)
parser.add_argument('--b-offset', type=int, default=0)
parser.add_argument('--padding', type=int, default=0)
parser.add_argument('--endpoint', type=str, default='')
parser.add_argument('--dataset', type=str, default='dataset')
opt = parser.parse_args()
args = vars(opt)
print('------------ Options -------------')
for k, v in sorted(args.items()):
print('%s: %s' % (str(k), str(v)))
print('-------------- End ----------------')
if not os.path.exists('./renders'):
os.mkdir('renders')
moduleNetwork = Network(opt.model).cuda()
tensorOutput = torch.FloatTensor()
index = 0
def recurse_two_frames(moduleNetwork, tensorOutput, a_np, b_np, frame_index, morph_index, step, depth=0):
print("generate {} {} {}".format(frame_index, morph_index, step))
tensorInputFirst = torch.FloatTensor(a_np)
tensorInputSecond = torch.FloatTensor(b_np)
process(moduleNetwork, tensorInputFirst, tensorInputSecond, tensorOutput)
middle_np = tensorOutput.clamp(0.0, 1.0).numpy()
if morph_index == frame_index:
print("frame {}, depth {}".format(frame_index, depth))
return middle_np
if morph_index > frame_index:
next_index = morph_index - step
next_a_np = a_np
next_b_np = middle_np
# print("next index: {} - {}".format(next_index, step))
else:
next_index = morph_index + step
next_a_np = middle_np
next_b_np = b_np
# print("next index: {} + {}".format(next_index, step))
return recurse_two_frames(moduleNetwork, tensorOutput, next_a_np, next_b_np, frame_index, next_index, step/2, depth+1)
def recurse_videos(moduleNetwork, tensorOutput, a, b, a_offset, b_offset, count, step, frame_index, dilate):
global index
index += 1
if (index % 10) == 0:
print("{}...".format(index))
step /= 2
a_fn = os.path.join(a, "frame_{:0>5}.png".format(int(frame_index + a_offset)))
b_fn = os.path.join(b, "frame_{:0>5}.png".format(int(frame_index + b_offset)))
print("{} => {}".format(a_fn, b_fn))
if not os.path.exists(a_fn) or not os.path.exists(b_fn):
return [ None ]
a_np = load_image(a_fn)
b_np = load_image(b_fn)
img_np = recurse_two_frames(moduleNetwork, tensorOutput, a_np, b_np, frame_index, count / 2, count / 4)
if step < 2 * dilate:
return [ img_np ]
else:
left = recurse_videos(moduleNetwork, tensorOutput, a, b, a_offset, b_offset, count, step, frame_index - (step/2), dilate)
right = recurse_videos(moduleNetwork, tensorOutput, a, b, a_offset, b_offset, count, step, frame_index + (step/2), dilate)
return left + [ img_np ] + right
def process_two_videos(moduleNetwork, tensorOutput, a, b, a_offset, b_offset, steps, dilate):
steps *= dilate
return recurse_videos(moduleNetwork, tensorOutput, a, b, a_offset, b_offset, steps, steps, steps/2, dilate)
def average_two_videos(moduleNetwork, tensorOutput, a, b, a_offset, b_offset, steps, dilate):
global index
index += 1
if (index % 10) == 0:
print("{}...".format(index))
frames = []
steps *= dilate
for i in range(1, steps * dilate + 1, dilate):
a_fn = os.path.join(a, "frame_{:0>5}.png".format(int(i + a_offset)))
b_fn = os.path.join(b, "frame_{:0>5}.png".format(int(i + b_offset)))
print("{} => {}".format(a_fn, b_fn))
if not os.path.exists(a_fn) and not os.path.exists(b_fn):
continue
a_np = load_image(a_fn)
b_np = load_image(b_fn)
tensorInputFirst = torch.FloatTensor(a_np)
tensorInputSecond = torch.FloatTensor(b_np)
process(moduleNetwork, tensorInputFirst, tensorInputSecond, tensorOutput)
middle_np = tensorOutput.clamp(0.0, 1.0).numpy()
frames.append(middle_np)
return frames
def process_tree(moduleNetwork, a, b, tensorOutput, steps, dilate):
global index
index += 1
if (index % 10) == 0:
print("{}...".format(index))
tensorInputFirst = torch.FloatTensor(a)
tensorInputSecond = torch.FloatTensor(b)
process(moduleNetwork, tensorInputFirst, tensorInputSecond, tensorOutput)
middle_np = tensorOutput.clamp(0.0, 1.0).numpy()
if steps < 4 * dilate:
return [middle_np]
else:
left = process_tree(moduleNetwork, a, middle_np, tensorOutput, steps / 2, dilate)
right = process_tree(moduleNetwork, middle_np, b, tensorOutput, steps / 2, dilate)
return left + [middle_np] + right
def smooth_frames(moduleNetwork, tensorOutput, frames, smooth):
if not smooth:
return frames
print("smoothing every other frame")
firstFrame = frames[0]
nextFrame = None
new_frames = [firstFrame]
for i in range(1, len(frames)-2, 2):
firstFrame = frames[i]
nextFrame = frames[i+2]
tensorInputFirst = torch.FloatTensor(firstFrame)
tensorInputSecond = torch.FloatTensor(nextFrame)
process(moduleNetwork, tensorInputFirst, tensorInputSecond, tensorOutput)
middle_np = tensorOutput.clamp(0.0, 1.0).numpy()
new_frames += [firstFrame]
new_frames += [middle_np]
if nextFrame is not None:
new_frames += [nextFrame]
new_frames += [frames[len(frames)-1]]
return new_frames
def dilate_frames(moduleNetwork, tensorOutput, frames, dilate):
if dilate < 2:
return frames
print("dilating by a factor of {}".format(dilate))
new_frames = []
nextFrame = frames[0]
for i in range(1, len(frames)):
firstFrame = nextFrame
nextFrame = frames[i]
new_frames += [firstFrame]
new_frames += process_tree(moduleNetwork, firstFrame, nextFrame, tensorOutput, dilate, 1)
new_frames += [nextFrame]
return new_frames
def store_frames(frames, outputPath, opt, inputFirst=None, inputSecond=None):
if not outputPath.endswith('.mp4'):
outputPath += '.mp4'
print('writing {}'.format(outputPath))
print('frames: {}'.format(len(frames)))
writer = FFMPEG_VideoWriter(outputPath, (1024, 512), FPS)
if inputFirst is not None:
writer.write_frame(tensor_to_image(inputFirst))
if opt.padding:
pad_frames(writer, opt.first, max(0, opt.offset_a - opt.padding * FPS), opt.offset_a)
for frame in frames:
if frame is not None:
writer.write_frame(tensor_to_image(frame))
if inputSecond is not None:
writer.write_frame(tensor_to_image(inputSecond))
if opt.padding:
pad_frames(writer, opt.second, opt.offset_b + len(frames), opt.offset_b + len(frames) + opt.padding * FPS)
writer.close()
if opt.endpoint != '':
call(["curl",
"-X", "POST",
"-F", "module=morph",
"-F", "activity=morph",
"-F", "generated=true",
"-F", "dataset=" + opt.dataset,
"-F", "datatype=video",
"-F", "should_relay=true",
"-F", "file=@" + outputPath,
opt.endpoint
])
def pad_frames(writer, base_path, start, end):
for index in range(start, end):
fn = os.path.join(base_path, "frame_{:0>5}.png".format(int(index + start)))
if os.path.exists(fn):
img = PIL.Image.open(fn)
writer.write_frame(img)
def tensor_to_image(np_val):
return (numpy.rollaxis(np_val, 0, 3)[:,:,::-1] * 255.0).astype(numpy.uint8)
def load_image(path):
return numpy.rollaxis(numpy.asarray(PIL.Image.open(path))[:,:,::-1], 2, 0).astype(numpy.float32) / 255.0
def load_image_tensor(path):
return torch.FloatTensor(load_image(path))
if opt.video and opt.video_out:
reader = FFMPEG_VideoReader(opt.video, False)
writer = FFMPEG_VideoWriter(opt.video_out, reader.size, reader.fps*2)
reader.initialize()
nextFrame = reader.read_frame()
for x in range(0, reader.nframes):
firstFrame = nextFrame
nextFrame = reader.read_frame()
tensorInputFirst = torch.FloatTensor(numpy.rollaxis(firstFrame[:,:,::-1], 2, 0) / 255.0)
tensorInputSecond = torch.FloatTensor(numpy.rollaxis(nextFrame[:,:,::-1], 2, 0) / 255.0)
process(moduleNetwork, tensorInputFirst, tensorInputSecond, tensorOutput)
writer.write_frame(firstFrame)
writer.write_frame((numpy.rollaxis(tensorOutput.clamp(0.0, 1.0).numpy(), 0, 3)[:,:,::-1] * 255.0).astype(numpy.uint8))
writer.write_frame(nextFrame)
writer.close()
elif opt.mix_videos:
print("morph two videos...")
outputPath = './renders/' + opt.video_out
frames = process_two_videos(moduleNetwork, tensorOutput, opt.first, opt.second, opt.a_offset, opt.b_offset, opt.steps, opt.dilate)
frames = smooth_frames(moduleNetwork, tensorOutput, frames, opt.smooth)
frames = dilate_frames(moduleNetwork, tensorOutput, frames, opt.dilate)
store_frames(frames, outputPath, opt)
elif opt.average_videos:
print("average two videos...")
outputPath = './renders/' + opt.video_out
frames = average_two_videos(moduleNetwork, tensorOutput, opt.first, opt.second, opt.a_offset, opt.b_offset, opt.steps, opt.dilate)
frames = smooth_frames(moduleNetwork, tensorOutput, frames, opt.smooth)
frames = dilate_frames(moduleNetwork, tensorOutput, frames, opt.dilate)
store_frames(frames, outputPath, opt)
elif opt.steps == 0:
print("generate single morphed image...")
tensorInputFirst = load_image_tensor(opt.first)
tensorInputSecond = load_image_tensor(opt.second)
process(moduleNetwork, tensorInputFirst, tensorInputSecond, tensorOutput)
PIL.Image.fromarray((numpy.rollaxis(tensorOutput.clamp(0.0, 1.0).numpy(), 0, 3)[:,:,::-1] * 255.0).astype(numpy.uint8)).save(opt.out)
elif opt.mix_images:
print("morph two video frames...")
inputFirst = load_image(os.path.join(opt.first, "frame_{:0>5}.png".format(opt.a_offset+1)))
inputSecond = load_image(os.path.join(opt.second, "frame_{:0>5}.png".format(opt.b_offset+1)))
outputPath = './renders/' + opt.video_out
frames = process_tree(moduleNetwork, inputFirst, inputSecond, tensorOutput, opt.steps * opt.dilate, opt.dilate)
frames = smooth_frames(moduleNetwork, tensorOutput, frames, opt.smooth)
print("dilate... {}".format(opt.dilate))
frames = dilate_frames(moduleNetwork, tensorOutput, frames, opt.dilate)
store_frames(frames, outputPath, inputFirst, inputSecond, opt)
else:
print("morph two images...")
inputFirst = load_image(opt.first)
inputSecond = load_image(opt.second)
outputPath = './renders/' + opt.video_out
frames = process_tree(moduleNetwork, inputFirst, inputSecond, tensorOutput, opt.steps * opt.dilate, opt.dilate)
frames = smooth_frames(moduleNetwork, tensorOutput, frames, opt.smooth)
print("dilate... {}".format(opt.dilate))
frames = dilate_frames(moduleNetwork, tensorOutput, frames, opt.dilate)
store_frames(frames, outputPath, opt, inputFirst, inputSecond)
|
