generated from eliahuhorwitz/Academic-project-page-template
-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathindex.html
630 lines (577 loc) · 31.1 KB
/
index.html
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
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<!-- Meta tags for social media banners, these should be filled in appropriatly as they are your "business card" -->
<!-- Replace the content tag with appropriate information -->
<meta name="description" content="DESCRIPTION META TAG">
<meta property="og:title" content="SOCIAL MEDIA TITLE TAG"/>
<meta property="og:description" content="SOCIAL MEDIA DESCRIPTION TAG TAG"/>
<meta property="og:url" content="URL OF THE WEBSITE"/>
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X630-->
<meta property="og:image" content="static/image/your_banner_image.png" />
<meta property="og:image:width" content="1200"/>
<meta property="og:image:height" content="630"/>
<meta name="twitter:title" content="TWITTER BANNER TITLE META TAG">
<meta name="twitter:description" content="TWITTER BANNER DESCRIPTION META TAG">
<!-- Path to banner image, should be in the path listed below. Optimal dimenssions are 1200X600-->
<meta name="twitter:image" content="static/images/your_twitter_banner_image.png">
<meta name="twitter:card" content="summary_large_image">
<!-- Keywords for your paper to be indexed by-->
<meta name="keywords" content="KEYWORDS SHOULD BE PLACED HERE">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Diffusion as Shader: 3D-aware Video Diffusion for Versatile Video Generation Control</title>
<link rel="icon" type="image/x-icon" href="static/images/favicon.ico">
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="static/css/bulma.min.css">
<link rel="stylesheet" href="static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="static/css/bulma-slider.min.css">
<link rel="stylesheet" href="static/css/fontawesome.all.min.css">
<link rel="stylesheet"
href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="static/css/index.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script src="https://documentcloud.adobe.com/view-sdk/main.js"></script>
<script defer src="static/js/fontawesome.all.min.js"></script>
<script src="static/js/bulma-carousel.min.js"></script>
<script src="static/js/bulma-slider.min.js"></script>
<script src="static/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">
Diffusion as Shader: 3D-aware Video Diffusion for Versatile Video Generation Control
</h1>
<div class="is-size-5 publication-authors">
<!-- Paper authors -->
<span class="author-block">
<a href="https://scholar.google.com/citations?user=Y8AU3RkAAAAJ&hl" target="_blank">Zekai Gu<sup>1</sup></a>,</span>
<span class="author-block">
<a href target="_blank">Rui Yan<sup>2</sup></a>,</span>
<span class="author-block">
<a href="https://github.com/jiah-cloud" target="_blank">Jiahao Lu<sup>1</sup></a>,</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=8eTLCkwAAAAJ&hl" target="_blank">Peng Li<sup>1</sup></a>,</span>
<span class="author-block">
<a href="https://frank-zy-dou.github.io/" target="_blank">Zhiyang Dou<sup>3</sup></a>,</span>
<span class="author-block">
<a href="https://scholar.google.com.sg/citations?user=XdahAuoAAAAJ" target="_blank">Chenyang Si<sup>4</sup></a>,</span>
<span class="author-block">
<a href="https://dongzhenwhu.github.io/index.html" target="_blank">Zhen Dong<sup>5</sup></a>,</span>
<span class="author-block">
<a href="https://openreview.net/profile?id=~Qifeng_Liu1" target="_blank">Qifeng Liu<sup>1</sup></a>,</span>
<span class="author-block">
<a href="https://clinplayer.github.io/" target="_blank">Cheng Lin<sup>3</sup></a>,</span>
<span class="author-block">
<a href="https://liuziwei7.github.io/" target="_blank">Ziwei Liu<sup>4</sup></a>,</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=28shvv0AAAAJ&hl" target="_blank">Wenping Wang<sup>6</sup></a>,</span>
<span class="author-block">
<a href="https://liuyuan-pal.github.io/" target="_blank">Yuan Liu<sup>1</sup></a></span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">
<sup>1</sup>The Hong Kong University of Science and Technology,
<sup>2</sup>Zhejiang University,
<sup>3</sup>The University of Hong Kong,<br>
<sup>4</sup>Nanyang Technological University
<sup>5</sup>Wuhan University
<sup>6</sup>Texas A&M University
</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<!-- Arxiv PDF link -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2501.03847.pdf" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper</span>
</a>
</span>
<!-- Supplementary PDF link -->
<!-- <span class="link-block">
<a href="static/pdfs/supplementary_material.pdf" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Supplementary</span>
</a>
</span> -->
<!-- Github link -->
<span class="link-block">
<a href="https://github.com/IGL-HKUST/DiffusionAsShader" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>Code</span>
</a>
</span>
<!-- ArXiv abstract Link -->
<span class="link-block">
<a href="http://arxiv.org/abs/2501.03847" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span>
<!-- Gallery Link -->
<span class="link-block">
<a href="more_results.html" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-images"></i>
</span>
<span>Gallery</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Teaser image -->
<section class="hero teaser">
<div class="container is-max-desktop">
<div class="hero-body">
<div style="margin-top: 15px;"></div>
<h2 class="subtitle has-text-centered">
We present <strong>Diffusion as Shader (DaS)</strong>, a versatile video generation control model <br>for all the following tasks.
</h2>
<div style="margin-top: 30px;"></div>
<video poster="" id="teaser" autoplay muted loop controlsList="nodownload" playsinline>
<!-- Your video file here -->
<source src="static/videos/teaser.mp4"
type="video/mp4">
</video>
</div>
</div>
</section>
<!-- Paper abstract -->
<section class="section hero is-light">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Abstract</h2>
<div class="content has-text-justified">
<p>
Diffusion models have demonstrated impressive performance in generating high-quality videos from text prompts or images. However, precise control over the video generation process—such as camera manipulation or content editing—remains a significant challenge. Existing methods for controlled video generation are typically limited to a single control type, lacking the flexibility to handle diverse control demands. In this paper, we introduce Diffusion as Shader (DaS), a novel approach that supports multiple video control tasks within a unified architecture. Our key insight is that achieving versatile video control necessitates leveraging 3D control signals, as videos are fundamentally 2D renderings of dynamic 3D content. Unlike prior methods limited to 2D control signals, DaS leverages 3D tracking videos as control inputs, making the video diffusion process inherently 3D-aware.
This innovation allows DaS to achieve a wide range of video controls by simply manipulating the 3D tracking videos. A further advantage of using 3D tracking videos is their ability to effectively link frames, significantly enhancing the temporal consistency of the generated videos. With just 3 days of fine-tuning on 8 H800 GPUs using less than 10k videos, DaS demonstrates strong control capabilities across diverse tasks, including mesh-to-video generation, camera control, motion transfer, and object manipulation.
</p>
</div>
</div>
</div>
</div>
</section>
<!-- End paper abstract -->
<!-- Object Manipulation Video carousel -->
<section class="hero is-small">
<div class="hero-body">
<div class="container" style="max-width: 1000px;">
<div style="margin-top: 50px;"></div>
<h2 class="title is-3 has-text-centered" style="font-size: 2em;">
Object Manipulation
<a href="more_results.html#object-manipulation" class="button is-small is-rounded is-dark" style="margin-left: 10px;">
More Results
</a>
</h2>
<div style="margin-top: 50px;"></div>
<div style="display: grid; grid-template-columns: repeat(2, 1fr); gap: 20px;">
<!-- 第一组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<img src="static/videos/manip/images/00001.png" alt="Object 1" style="width: 98%;">
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/manip/results/00001.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- 第二组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<img src="static/videos/manip/images/00002.png" alt="Object 2" style="width: 98%;">
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/manip/results/00002.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- 第3组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<img src="static/videos/manip/images/1.png" alt="Object 1" style="width: 98%;">
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/manip/results/1.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- 第4组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<img src="static/videos/manip/images/3.png" alt="Object 2" style="width: 98%;">
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/manip/results/3.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
<div style="margin-top: 10px;"></div>
<h2 class="content has-text-justified">
DaS can generate a video to manipulate a specific object.
Given an image, we estimate the depth map using Depth Pro or MoGE and segment out the object using SAM.
Then, we are able to manipulate the point cloud of the object to construct a 3D tracking video for object manipulation video generation.
</h2>
</div>
</div>
</section>
<!-- End Object Manipulation Video carousel -->
<!-- Video list -->
<section class="hero is-small">
<div class="hero-body">
<div class="container" style="max-width: 1000px;">
<div style="margin-top: 50px;"></div>
<h2 class="title is-3 has-text-centered" style="font-size: 2em;">
Animating meshes to videos
<a href="more_results.html#mesh-to-video" class="button is-small is-rounded is-dark" style="margin-left: 10px;">
More Results
</a>
</h2>
<div style="margin-top: 50px;"></div>
<div id="results-list" class="results-list">
<div style="width: 80%; margin: 0 auto;">
<div class="item item-video5" style="position: relative;">
<video poster="" id="video5" autoplay muted loop height="100%" controlsList="nodownload" playsinline>
<source src="static/videos/animation/m2vmix.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="columns is-2">
<div class="column">
<div class="item item-video1" style="position: relative;">
<div style="margin-top: 20px;"></div>
<h3 class="title is-5 has-text-centered">Rendering style: A muscular man</h3>
<video poster="" id="video1" autoplay muted loop height="80%" controlsList="nodownload" playsinline>
<source src="static/videos/animation/m2v1.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="column">
<div class="item item-video2" style="position: relative;">
<div style="margin-top: 20px;"></div>
<h3 class="title is-5 has-text-centered">Rendering style: Man in a suit</h3>
<video poster="" id="video2" autoplay muted loop height="80%" controlsList="nodownload" playsinline>
<source src="static/videos/animation/m2v2.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
</div>
<div style="margin-top: 20px;"></div>
<h2 class="content has-text-justified">
DaS enables the creation of visually appealing, high-quality videos from simple animated meshes.
While many Computer Graphics (CG) software tools provide basic 3D models and motion templates to generate animated meshes,
these outputs are often simplistic and lack the detailed appearance and geometry needed for high-quality animations.
Starting with these simple animated meshes, we generate an initial visually appealing frame using a depth-to-image FLUX model.
We then produce 3D tracking videos from the animated meshes, which, when combined with the generated first frame, guide DaS to transform the basic meshes into visually rich and appealing videos.
</h2>
</video>
</div>
</div>
</section>
<!-- End video carousel -->
<!-- Camera Control Video carousel -->
<section class="hero is-small">
<div class="hero-body">
<div class="container" style="max-width: 1000px;">
<div style="margin-top: 50px;"></div>
<h2 class="title is-3 has-text-centered" style="font-size: 2em;">
Camera Control
<a href="more_results.html#camera-control" class="button is-small is-rounded is-dark" style="margin-left: 10px;">
More Results
</a>
</h2>
<div style="margin-top: 50px;"></div>
<div style="display: flex; justify-content: space-between;">
<div style="width: 50%; text-align: center; padding: 10px;">
<img src="static/videos/scenes/pose.png" alt="Camera Control" style="width: 100%;">
</div>
<div style="width:48%; display: flex; flex-wrap: wrap; justify-content: space-between;">
<div class="item item-video1" style="width: 49%; height: 40%;">
<h3 class="title is-4 has-text-centered">Moving Up</h3>
<video poster="" id="video1" autoplay muted loop height="90%">
<source src="static/videos/scenes/c1.mp4" type="video/mp4">
</video>
</div>
<div class="item item-video2" style="width: 49%; height: 40%;">
<h3 class="title is-4 has-text-centered">Moving Down</h3>
<video poster="" id="video2" autoplay muted loop height="90%">
<source src="static/videos/scenes/c4.mp4" type="video/mp4">
</video>
</div>
<div class="item item-video3" style="width: 49%; height: 40%;">
<h3 class="title is-4 has-text-centered">Moving Right</h3>
<video poster="" id="video3" autoplay muted loop height="90%">
<source src="static/videos/scenes/c2.mp4" type="video/mp4">
</video>
</div>
<div class="item item-video4" style="width: 49%; height: 40%;">
<h3 class="title is-4 has-text-centered">Moving Left</h3>
<video poster="" id="video4" autoplay muted loop height="90%">
<source src="static/videos/scenes/c3.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
<div style="margin-top: 10px;"></div>
<h2 class="content has-text-justified">
DaS significantly enhances 3D awareness by incorporating 3D tracking videos for precise camera control.
To generate videos with a specific camera trajectory, we first estimate the depth map of the initial frame using Depth Pro and convert it into colored 3D points.
These points are then projected onto the given camera trajectory, constructing a 3D tracking video that enables DaS to control camera movements with high 3D accuracy.
</h2>
<div style="margin-top: 50px;"></div>
<div style="display: flex; justify-content: space-between; align-items: center; font-size: 1.5em;">
<div style="width: 25%; text-align: center;">Moving Right</div>
<div style="width: 25%; text-align: center;">Moving Left</div>
<div style="width: 25%; text-align: center;">Moving Up</div>
<div style="width: 25%; text-align: center;">Moving Down</div>
</div>
<video poster="" id="camctrl" autoplay muted loop height="90%">
<source src="static/videos/scenes/camctrl.mp4" type="video/mp4">
</video>
<div style="display: flex; flex-direction: column; align-items: center;">
<div style="margin-top: 20px;"></div>
<!-- 添加 Spiral 标题 -->
<h2 class="title is-3 has-text-centered" style="font-size: 1.5em; margin-bottom: 20px;">Spiral
<a href="more_results.html#camera-control" class="button is-small is-rounded is-dark" style="margin-left: 10px;">
More Results
</a>
</h2>
<!-- 使用 CSS Grid 布局，定义两行两列 -->
<div style="display: grid; grid-template-columns: repeat(2, 1fr); grid-template-rows: repeat(2, 1fr); gap: 20px;">
<!-- 第一组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<img src="static/videos/spiral/images/00022.png" alt="Scene 1" style="width: 98%;">
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/spiral/results/00022.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- 第二组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<img src="static/videos/spiral/images/00014.png" alt="Scene 2" style="width: 98%;">
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/spiral/results/00014.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- 第三组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/spiral/images/s1.mp4" type="video/mp4">
</video>
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/spiral/results/1.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- 第四组：原图像和生成视频 -->
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-image" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/spiral/images/s2.mp4" type="video/mp4">
</video>
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/spiral/results/2.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
</div>
</section>
<!-- End Camera Control Video carousel -->
<!-- Motion Transfer -->
<section class="hero is-small">
<div class="hero-body">
<div class="container" style="max-width: 1000px;">
<div style="margin-top: 50px;"></div>
<h2 class="title is-3 has-text-centered" style="font-size: 2em;">
Motion Transfer
<a href="more_results.html#motion-transfer" class="button is-small is-rounded is-dark" style="margin-left: 10px;">
More Results
</a>
</h2>
<div style="margin-top: 50px;"></div>
<div style="display: flex; justify-content: space-between; align-items: center; font-size: 1em;">
<div style="width: 25%; text-align: center;">Source</div>
<div style="width: 25%; text-align: center;">Transferred1</div>
<div style="width: 25%; text-align: center;">Transferred2</div>
<div style="width: 25%; text-align: center;">Transferred3</div>
</div>
<div class="column is-12">
<video poster="" id="video1" autoplay muted loop height="100%">
<!-- Your video file here -->
<source src="static/videos/l1.mp4"
type="video/mp4">
</video>
<video poster="" id="video2" autoplay muted loop height="100%">
<!-- Your video file here -->
<source src="static/videos/l2.mp4"
type="video/mp4">
</video>
</div>
<div style="margin-top: 20px;"></div>
<div style="display: grid; grid-template-columns: repeat(2, 1fr); grid-template-rows: repeat(1, 1fr); gap: 15px;">
<!-- 3-4组视频 -->
<div style="display: flex; flex-direction: column; align-items: center;">
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/mr/gt/a3.mp4" type="video/mp4">
</video>
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/mr/gen/a3.mp4" type="video/mp4">
</video>
</div>
</div>
<div style="margin-top: 10px; font-style: italic; color: #666;">"A yellow fox runs on the grass"</div>
</div>
<div style="display: flex; flex-direction: column; align-items: center;">
<div class="video-group" style="display: flex; align-items: center; justify-content: center;">
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/mr/gt/a4.mp4" type="video/mp4">
</video>
</div>
<div class="item item-video" style="flex: 1;">
<video autoplay muted loop style="width: 98%;">
<source src="static/videos/mr/gen/a4.mp4" type="video/mp4">
</video>
</div>
</div>
<div style="margin-top: 10px; font-style: italic; color: #666;">
“A herd of bird-deer in a towering, wooded forest.”
</div>
</div>
</div>
<div style="margin-top: 10px;"></div>
<h2 class="content has-text-justified">
DaS also facilitates creating a new video by transferring motion from an existing source video.
First, we estimate the depth map of the source video’s first frame and apply the depth-to-image FLUX model to repaint the frame into a target appearance guided by text prompts.
Then, using SpatialTracker, we generate a 3D tracking video from the source video to serve as control signals.
Finally, the DaS model generates the target video by combining the edited first frame with the 3D tracking video.
</h2>
</div>
</div>
</section>
<script>
document.addEventListener('DOMContentLoaded', function() {
const motionTransferVideos = document.querySelectorAll('.video-group video');
const spiralVideos = document.querySelectorAll('.item-video video');
const allControlledVideos = [...motionTransferVideos, ...spiralVideos];
allControlledVideos.forEach(video => {
const source = video.querySelector('source');
if (source && source.src.includes('/mr/gt/')) {
video.playbackRate = 0.333;
video.addEventListener('loadedmetadata', function() {
this.playbackRate = 0.333;
});
}
video.addEventListener('click', function() {
if (this.paused) {
allControlledVideos.forEach(v => v.play());
} else {
allControlledVideos.forEach(v => v.pause());
}
});
video.addEventListener('ended', function() {
allControlledVideos.forEach(v => {
v.currentTime = 0;
v.play();
});
});
});
Promise.all([...allControlledVideos].map(video => {
return new Promise(resolve => {
video.addEventListener('loadedmetadata', resolve);
});
})).then(() => {
allControlledVideos.forEach(video => video.play());
});
});
</script>
<!--BibTex citation -->
<section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>@article{gu2025das,
title={Diffusion as Shader: 3D-aware Video Diffusion for Versatile Video Generation Control},
author={Zekai Gu and Rui Yan and Jiahao Lu and Peng Li and Zhiyang Dou and Chenyang Si and Zhen Dong and Qifeng Liu and Cheng Lin and Ziwei Liu and Wenping Wang and Yuan Liu},
year={2025},
journal={arXiv preprint arXiv:2501.03847}
}
</code></pre>
</div>
</section>
<!--End BibTex citation -->
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This page was built using the <a href="https://github.com/eliahuhorwitz/Academic-project-page-template" target="_blank">Academic Project Page Template</a> which was adopted from the <a href="https://nerfies.github.io" target="_blank">Nerfies</a> project page.
You are free to borrow the source code of this website, we just ask that you link back to this page in the footer. <br> This website is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
<!-- Statcounter tracking code -->
<!-- You can add a tracker to track page visits by creating an account at statcounter.com -->
<!-- End of Statcounter Code -->
</body>
</html>