-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
702 lines (642 loc) · 33.6 KB
/
Copy pathindex.html
File metadata and controls
702 lines (642 loc) · 33.6 KB
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
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
<!DOCTYPE html>
<html>
<!-- <link href="https://fonts.cdnfonts.com/css/caveat" rel="stylesheet">
<link href="https://fonts.googleapis.com/css?family=Handlee&v1" rel="stylesheet">
<style>
@import url('https://fonts.cdnfonts.com/css/caveat');
</style> -->
<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>SkyReels-A2</title>
<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>
<nav class="navbar" role="navigation" aria-label="main navigation">
<div class="navbar-brand">
<a role="button" class="navbar-burger" aria-label="menu" aria-expanded="false">
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
<span aria-hidden="true"></span>
</a>
</div>
<div class="navbar-menu">
<div class="navbar-start" style="flex-grow: 1; justify-content: center;">
<div class="navbar-item has-dropdown is-hoverable">
<a class="navbar-link">
More Research
</a>
<div class="navbar-dropdown">
<a class="navbar-item" href="https://skyworkai.github.io/skyreels-a1.github.io/">
SkyReels-A1
</a>
</a>
</div>
</div>
</div>
</div>
</nav>
<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-3 publication-title">SkyReels-A2: Compose Anything in Video Diffusion Transformers</h1>
<div class="is-size-5 publication-authors">
<!-- Paper authors -->
<span class="author-block">
<a href="https://scholar.google.com.hk/citations?hl=zh-CN&user=_43YnBcAAAAJ" target="_blank">Zhengcong Fei</a><sup>*</sup>,
</span>
<span class="author-block">
<a href="https://debangli.github.io/" target="_blank">Debang Li</a><sup>* †</sup>,
</span>
<span class="author-block">
<a href="https://scholar.google.com.hk/citations?hl=zh-CN&user=6D_nzucAAAAJ" target="_blank">Di Qiu</a><sup>* †</sup>,
</span>
<span class="author-block">
<a href="" target="_blank">Jiahua Wang</a>,
</span>
<span class="author-block">
<a href="" target="_blank">Yikun Dou</a>,
</span>
<span class="author-block">
<a href="" target="_blank">Rui Wang</a>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=bL_o7I8AAAAJ&hl=en" target="_blank">Jingtao Xu</a>,
</span>
<span class="author-block">
<a href="" target="_blank">Mingyuan Fan</a>,
</span>
<span class="author-block">
<a href="https://scholar.google.com.hk/citations?hl=zh-CN&user=HukWSw4AAAAJ&view_op=list_works&sortby=pubdate" target="_blank">Guibin Chen</a>,
</span>
<span class="author-block">
<a href="" target="_blank">Yang Li</a>,
</span>
<span class="author-block">
<a href="" target="_blank">Yahui Zhou</a>
</span>
</div>
<div class="is-size-5 publication-authors">
<p><span><sup>*</sup>Equal contribution,<sup> †</sup>Project lead</span></p>
<span class="author-block">Skywork AI, Kunlun Inc.</span>
</div>
<br>
<div class="column has-text-centered">
<div class="publication-links">
<!-- Arxiv PDF link -->
<span class="link-block">
<a href="https://arxiv.org/pdf/2504.02436" 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>
<span class="link-block">
<a href="https://github.com/SkyworkAI/SkyReels-A2"
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>
<!-- Supplementary PDF link -->
<span class="link-block">
<a href="https://huggingface.co/Skywork/SkyReels-A2"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa fa-smile"></i>
</span>
<span>Models</span>
</a>
</span>
<span class="link-block">
<a href="https://huggingface.co/datasets/Skywork/A2-Bench"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa fa-smile"></i>
</span>
<span>A2-Bench</span>
</a>
</span>
<span class="link-block">
<a href="https://huggingface.co/spaces/Skywork/SkyReels_A2_Bench"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa fa-smile"></i>
</span>
<span>Leaderboard</span>
</a>
</span>
<span class="link-block">
<a href="https://www.skyreels.ai"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fa fa-grin-tongue"></i>
</span>
<span>Demo (SkyReels)</span>
</a>
</span>
</span>
<!-- Github link -->
<!-- <span class="link-block">
<a href="https://github.com/omerbt/TokenFlow" 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> -->
<!-- Hugging Face Demo link with an image icon -->
<!-- <span class="link-block">
<a href="https://huggingface.co/spaces/weizmannscience/tokenflow" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<img src="static/images/hf.png" alt="Hugging Face Demo">
</span>
<span>Demo</span>
</a>
</span> -->
<!-- ArXiv abstract Link -->
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- Teaser video-->
<section class="hero teaser is-light" align="center">
<div class="container is-max-desktop">
<br>
<h2 class="title is-3">Overview of Generated Videos</h2>
<div class="hero-body">
<video poster="" id="tree" autoplay controls muted loop height="100%">
<!-- Your video here -->
<source src="static/videos/overall.mp4"
type="video/mp4">
</video>
<!-- <h2 class="subtitle has-text-centered">
Aliquam vitae elit ullamcorper tellus egestas pellentesque. Ut lacus tellus, maximus vel lectus at, placerat pretium mi. Maecenas dignissim tincidunt vestibulum. Sed consequat hendrerit nisl ut maximus.
</h2> -->
</div>
</div>
</section>
<!-- End teaser video -->
<!-- Paper abstract -->
<section class="hero is-small">
<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>
This paper presents \texttt{SkyReels-A2}, a controllable video generation framework capable of assembling arbitrary visual elements (e.g., characters, objects, backgrounds) into synthesized videos based on textual prompts while maintaining strict consistency with reference images for each element. We term this task \emph{elements-to-video (E2V)}, whose primary challenges lie in preserving per-element fidelity to references, ensuring coherent scene composition, and achieving natural outputs. To address these, we first design a comprehensive data pipeline to construct prompt-reference-video triplets for model training. Next, we propose a novel image-text joint embedding model to inject multi-element representations into the generative process, balancing element-specific consistency with global coherence and text alignment. We also optimize the inference pipeline for both speed and output stability. Moreover, we introduce a carefully curated benchmark for systematic evaluation, i.e, \texttt{A2 Bench}. Experiments demonstrate that our framework can generate diverse, high-quality videos with precise element control. \texttt{SkyReels-A2} is the first commercial-grade open-source model for \emph{E2V} generation, performing favorably against advanced commercial closed-source models. We anticipate \texttt{SkyReels-A2} will advance creative applications such as drama and virtual e-commerce, pushing the boundaries of controllable video generation.
</p>
</div>
</div>
</div>
</div>
</section>
<!-- End paper abstract -->
<!-- Paper poster -->
<section class="hero is-small is-light">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-">
<h2 class="title is-3">Method</h2>
<div class="content has-text-justified">
<!--
<table width="800" border="0">
<tbody>
<tr>
<td colspan="3">
<p>
We observe that the level of temporal consistency of a video is tightly related to the temporal consistency of its feature representation, as can be seen in the feature visualization below.
The features of a natural video have a shared, temporally consistent representation. When editing the video per frame, this consistency breaks. Our method ensures the same level of feature consistency as in the original video features.
</p>
</td>
</tr>
<tr>
<td style="font-size: 16px; text-align: center;">Original</td>
<td style="font-size: 16px; text-align: center;">Per Frame Editing</td>
<td style="font-size: 16px; text-align: center;">Ours</td>
</tr>
<tr class="video-row">
<td style="text-align: center;">
<a href="sm/assets/man_basket/input_fps30.mp4">
<video preload="auto"width="224" src="sm/assets/man_basket/input_fps30.mp4" autoplay loop controls muted/>
</a>
</td>
<td style="text-align: center;">
<a href="sm/assets/man_basket/Van-Gogh style portrait of a man spinning a basketball, oil painting, art by Van Gogh, 8k/pnp_per_frame_baseline_fps_30.mp4">
<video preload="auto"width="224" src="sm/assets/man_basket/Van-Gogh style portrait of a man spinning a basketball, oil painting, art by Van Gogh, 8k/pnp_per_frame_baseline_fps_30.mp4" autoplay loop controls muted/>
</a>
</td>
<td style="text-align: center;">
<a href="sm/assets/man_basket/Van-Gogh style portrait of a man spinning a basketball, oil painting, art by Van Gogh, 8k/result_fps30.mp4">
<video preload="auto"width="224" src="sm/assets/man_basket/Van-Gogh style portrait of a man spinning a basketball, oil painting, art by Van Gogh, 8k/result_fps_30.mp4" autoplay loop controls muted/>
</a>
</td>
</tr>
<tr class="video-row">
<td style="text-align: center;">
<a href="videos/pca/tokens_origvideo_30.mp4">
<video preload="auto"width="224" src="videos/pca/tokens_origvideo_30.mp4" autoplay loop controls muted/>
</a>
</td>
<td style="text-align: center;">
<a href="videos/pca/tokens_pnpvideo_30.mp4">
<video preload="auto"width="224" src="videos/pca/tokens_pnpvideo_30.mp4" autoplay loop controls muted/>
</a>
</td>
<td style="text-align: center;">
<a href="videos/pca/tokens_flowvideo_30.mp4">
<video preload="auto"width="224" src="videos/pca/tokens_flowvideo_30.mp4" autoplay loop controls muted/>
</a>
</td>
</tr>
<tr>
<td colspan="3">
<p style="margin-top: 20px; margin-bottom: -12px;">
Our key finding is that a temporally-consistent edit can be achieved by enforcing consistency on the internal diffusion features across frames during the editing process.
We achieve this by propagating a small set of edited features across frames, using the correspondences between the original video features.
Given an input video I, we invert each frame, extract its tokens (i.e., output features from the self-attention modules), and extract inter-frame feature correspondences using a nearest-neighbor (NN) search. At each denoising step:
</p>
<ol>
(I) We sample keyframes from the noisy video J_t and jointly edit them using an extended-attention block. The set of resulting edited tokens is T_base.</li>
<br>
(II) We propagate the edited tokens across the video according to the pre-computed correspondences of the original video features.</li>
</ol>
To denoise J_t, we feed each frame to the network and replace the generated tokens with the tokens obtained from the propagation step (II).
</td>
</tr>
</tbody>
</table> -->
<div>
<td colspan="3"><img src="static/images/framework.png" alt="" width="1000" /></td>
</div>
<p>Overview of SkyReels-A2 framework. Our approach initiates by encoding all reference images using two distinct branches. The first, termed the spatial feature branch (represented in red, top row), leverages a fine-grained VAE encoder to process per-composition images. The second, identified as the semantic feature branch (represented in red, bottom row), utilizes a CLIP vision encoder followed by an MLP projection to encode semantic references. Subsequently, the spatial features are concatenated with the noised video tokens along the channel dimension before being passed through the diffusion transformer blocks. Meanwhile, the semantic features extracted from the reference images are incorporated into the diffusion transformers via supplementary cross-attention layers, ensuring that the semantic context is effectively integrated during diffusion. </p>
</div>
</div>
</div>
</div>
</div>
</section>
<!--End paper poster -->
<section class="hero is-small"></section>
<div class="hero-body">
<!-- <div class="container is-max-desktop"> -->
<div class="columns is-centered has-text-centered">
<div class="column custom-width">
<h2 class="title is-3">Examples</h2>
<h3 class="title is-5">Human-Subjective-Background Tuples</h3>
<div class="content has-text-justified" align="center">
<div id="results-carousel" class="carousel results-carousel" align="center">
<div class="item item-video11">
<video preload="auto"poster="" id="video11" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/single-human/1.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video12">
<video preload="auto"poster="" id="video12" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/single-human/2.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video13">
<video preload="auto"poster="" id="video13" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/single-human/3.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video14">
<video preload="auto"poster="" id="video14" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/single-human/4.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video15">
<video preload="auto"poster="" id="video15" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/single-human/5.mp4"
type="video/mp4">
</video>
</div>
</div>
</div>
<p style="text-align: left; max-width: 1024px; margin-left: 100px;">SkyReels-A2 enables character, objectives, and background reference image to composite a natural videos.</p>
</div>
</div>
</div>
<!-- <p align="center"> .</p> -->
</section>
<section class="hero is-small"></section>
<div class="hero-body">
<!-- <div class="container is-max-desktop"> -->
<div class="columns is-centered has-text-centered">
<div class="column custom-width">
<h3 class="title is-5">Multiple-Human-Background Tuples</h3>
<div class="content has-text-justified" align="center">
<div id="results-carousel" class="carousel results-carousel" align="center">
<div class="item item-video1">
<video preload="auto"poster="" id="video1" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/multi-human/1.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video2">
<video preload="auto"poster="" id="video2" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/multi-human/2.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video3">
<video preload="auto"poster="" id="video3" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/multi-human/3.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video4">
<video preload="auto"poster="" id="video4" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/multi-human/4.mp4"
type="video/mp4">
</video>
</div>
</div>
</div>
<p style="text-align: left; max-width: 1024px; margin-left: 100px;">SkyReels-A2 also support multiple-human reference composition, creating high-quality interactive videos.</p>
</div>
</div>
</div>
<!-- <p align="center">Qualitative comparison of customized video generation with both subjects and motions. <br>Without guidance from additional videos, our method significantly outperforms in terms of concept combination.</p> -->
</section>
<section class="hero is-small"></section>
<div class="hero-body">
<!-- <div class="container is-max-desktop"> -->
<div class="columns is-centered has-text-centered">
<div class="column custom-width">
<h3 class="title is-5">Virtual E-commerce</h3>
<div class="content has-text-justified" align="center">
<div id="results-carousel" class="carousel results-carousel" align="center">
<div class="item item-video1">
<video preload="auto"poster="" id="video1" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/shopping/1.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video2">
<video preload="auto"poster="" id="video2" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/shopping/2.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video3">
<video preload="auto"poster="" id="video3" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/shopping/3.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video4">
<video preload="auto"poster="" id="video4" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/shopping/4.mp4"
type="video/mp4">
</video>
</div>
</div>
</div>
<p style="text-align: left; max-width: 1024px; margin-left: 100px; ">An important application of SkyReels-A2 is the ability to generate corresponding recommendation scenarios based on the anchor and product images.</p>
</div>
</div>
</div>
<!-- <p align="center">Qualitative comparison of customized video generation with both subjects and motions. <br>Without guidance from additional videos, our method significantly outperforms in terms of concept combination.</p> -->
</section>
<section class="hero is-small"></section>
<div class="hero-body">
<!-- <div class="container is-max-desktop"> -->
<div class="columns is-centered has-text-centered">
<div class="column custom-width">
<h3 class="title is-5">AI MV</h3>
<div class="content has-text-justified" align="center">
<div id="results-carousel" class="carousel results-carousel" align="center">
<div class="item item-video1">
<video preload="auto"poster="" id="video1" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/mv/1.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video2">
<video preload="auto"poster="" id="video2" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/mv/2.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video3">
<video preload="auto"poster="" id="video3" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/mv/3.mp4"
type="video/mp4">
</video>
</div>
<div class="item item-video4">
<video preload="auto"poster="" id="video4" autoplay controls muted loop width="480px">
<!-- Your video preload="auto"file here -->
<source src="static/videos/mv/4.mp4"
type="video/mp4">
</video>
</div>
</div>
</div>
<p style="text-align: left; max-width: 1024px; margin-left: 100px;">SkyReels-A2 also highlights the effectiveness in building music multimedia creation scene.</p>
</div>
</div>
</div>
<!-- <p align="center">Qualitative comparison of customized video generation with both subjects and motions. <br>Without guidance from additional videos, our method significantly outperforms in terms of concept combination.</p> -->
</section>
<!-- Paper poster -->
<section class="hero is-small is-light">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-">
<h2 class="title is-3">Data Construction Pipeline</h2>
<div class="content has-text-justified">
<div>
<td colspan="3"><img src="static/images/datapipeline.png" alt="" width="1000" /></td>
</div>
<p>The pipeline begins with preprocessing, where raw videos are filtered by resolution, labels, types, and sources, followed by temporal segmentation based on key-frames. Next, a proprietary multi-expert video captioning model generates both holistic descriptions for video clips and structured concept annotations. Subsequently, detection and segmentation models extract visual elements (e.g., humans, objects, environments). To mitigate duplication, reference images are retrieved from other clips based on clip/facial similarity score. Further refinement includes face detection and human parsing to obtain facial/attire elements. Finally, the extracted elements are matched with structured descriptions to form training triplets (visual elements, video clips, and textual captions)</p>
</div>
</div>
</div>
</div>
</div>
</section>
<!--End paper poster -->
<!-- Paper poster -->
<section class="hero is-small">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-">
<h2 class="title is-3">A2 Bench</h2>
<div class="content has-text-justified">
<div>
<td colspan="3"><img src="static/images/bench.png" alt="" width="1000" /></td>
</div>
<p>The dimensions covered in A2-Bench. Our evaluation consider both automatic metrics and user study, meantime, it covers multiple perspectives that precisely reflects the quality of elements-to-video (E2V) task.</p>
</div>
</div>
</div>
</div>
</div>
</section>
<!--End paper poster -->
<section class="hero is-small is-light">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column custom-width">
<h2 class="title is-3">References</h2>
<div class="content has-text-justified">
<p>
<a name="ConcepMaster" id="ConcepMaster"></a>
[1] Huang Y, Yuan Z, Liu Q, et al. ConceptMaster: Multi-Concept Video Customization on Diffusion Transformer Models Without Test-Time Tuning[J]. arXiv preprint arXiv:2501.04698, 2025.
</p>
<p>
<a name="Phantom" id="Phantom"></a>
[2] Liu L, Ma T, Li B, et al. Phantom: Subject-consistent video generation via cross-modal alignment[J]. arXiv preprint arXiv:2502.11079, 2025.
</p>
<p>
<a name="wanx" id="wanx"></a>
[3] Wang A, Ai B, Wen B, et al. Wan: Open and Advanced Large-Scale Video Generative Models[J]. arXiv preprint arXiv:2503.20314, 2025.
</p>
</div>
</div>
</div>
</div>
</div>
</section>
<!--BibTex citation -->
<!-- <section class="section" id="BibTeX">
<div class="container is-max-desktop content">
<h2 class="title">BibTeX</h2>
<pre><code>
@article{wu2024customcrafter,
title={CustomCrafter: Customized Video Generation with Preserving Motion and Concept Composition Abilities},
author={Wu, Tao and Zhang, Yong and Wang, Xintao and Zhou, Xianpan and Zheng, Guangcong and Qi, Zhongang and Shan, Ying and Li, Xi},
journal={arXiv preprint arXiv:2408.13239},
year={2024}
}
</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>.
<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 -->
<script>
window.addEventListener('DOMContentLoaded', (event) => {
const videoWrappers = document.querySelectorAll('.video-wrapper');
videoWrappers.forEach(wrapper => {
const defaultVideo = wrapper.querySelector('.default-video');
const aspectRatio = defaultVideo.videoWidth / defaultVideo.videoHeight;
const height = wrapper.offsetWidth / aspectRatio;
wrapper.style.height = `${height}px`;
wrapper.addEventListener('mouseenter', () => {
defaultVideo.pause();
hoverVideo.play();
});
wrapper.addEventListener('mouseleave', () => {
defaultVideo.play();
hoverVideo.pause();
});
});
});
$(document).ready(function() {
var carouselItems = $('.carousel .item');
var numItems = carouselItems.length;
var numVideos = 5;
var currentIndex = 0;
$('.carousel').on('click', function() {
currentIndex++;
if (currentIndex + numVideos <= numItems) {
carouselItems.removeClass('active');
carouselItems.slice(currentIndex, currentIndex + numVideos).addClass('active');
} else {
currentIndex = 0;
carouselItems.removeClass('active');
carouselItems.slice(currentIndex, currentIndex + numVideos).addClass('active');
}
});
carouselItems.slice(currentIndex, currentIndex + numVideos).addClass('active');
});
</script>
</body>
</html>