-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
850 lines (760 loc) · 43 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
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
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="VBench: Comprehensive Benchmark Suite for Video Generative Models">
<meta name="keywords" content="Evaluation, Benchmark, Prompts, Dataset, Video Generation, Text-to-Video, Stable Diffusion, AIGC, Image Generation, MMLab, Shanghai AI Lab">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>VBench: Comprehensive Benchmark Suite for Video Generative Models</title>
<!-- Global site tag (gtag.js) - Google Analytics-->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-Y5ZVQZ7NHC"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag() {
dataLayer.push(arguments);
}
gtag('js', new Date());
gtag('config', 'G-Y5ZVQZ7NHC');
</script>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="https://cdn.datatables.net/1.13.6/css/jquery.dataTables.min.css">
<link rel="stylesheet" href="./assets/vbench/css/bulma.min.css">
<link rel="stylesheet" href="./assets/vbench/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./assets/vbench/css/bulma-slider.min.css">
<link rel="stylesheet" href="./assets/vbench/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="./assets/vbench/css/index.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./assets/vbench/js/fontawesome.all.min.js"></script>
<script src="./assets/vbench/js/bulma-carousel.min.js"></script>
<script src="./assets/vbench/js/bulma-slider.min.js"></script>
<script src="./assets/vbench/js/index.js"></script>
</head>
<body>
<!-- title -->
<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"><span style="color:#B9770E; font-weight: bold; font-style: italic">VBench</span> : Comprehensive Benchmark<br>Suite for Video Generative Models</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">CVPR 2024 Highlight</span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">
<span class="author-block">
<a href="https://ziqihuangg.github.io" target="_blank">Ziqi Huang</a><sup>1*</sup>,
</span>
<span class="author-block">
<a href="https://github.com/yinanhe" target="_blank">Yinan He</a><sup>2*</sup>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=iH0Aq0YAAAAJ&hl=zh-CN" target="_blank">Jiashuo Yu</a><sup>2*</sup>,
</span>
<span class="author-block">
<a href="https://github.com/zhangfan-p" target="_blank">Fan Zhang</a><sup>2*</sup>,
</span>
<br>
<span class="author-block">
<a href="https://chenyangsi.top/" target="_blank">Chenyang Si</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://yumingj.github.io/" target="_blank">Yuming Jiang</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://zhangyuanhan-ai.github.io/" target="_blank">Yuanhan Zhang</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://tianxingwu.github.io/" target="_blank">Tianxing Wu</a><sup>1</sup>,
</span>
<span class="author-block">
<a target="_blank">Qingyang Jin</a><sup>1</sup>,
</span>
<span class="author-block">
<a href="https://nattapolchan.github.io/me" target="_blank">Nattapol Chanpaisit</a><sup>1</sup>,
</span>
<br>
<span class="author-block">
<a href="https://wyhsirius.github.io/" target="_blank">Yaohui Wang</a><sup>2</sup>,
</span>
<span class="author-block">
<a href="https://scholar.google.com/citations?user=3fWSC8YAAAAJ" target="_blank">Xinyuan Chen</a><sup>2</sup>,
</span>
<span class="author-block">
<a href="https://wanglimin.github.io" target="_blank">Limin Wang</a><sup>4,2</sup>,
</span>
<span class="author-block">
<a href="http://dahua.site/" target="_blank">Dahua Lin</a><sup>2,3†</sup>,
</span>
<span class="author-block">
<a href="http://mmlab.siat.ac.cn/yuqiao/index.html" target="_blank">Yu Qiao</a><sup>2†</sup>,
</span>
<span class="author-block">
<a href="https://liuziwei7.github.io/" target="_blank">Ziwei Liu</a><sup>1†</sup>
</span>
</div>
<!-- <br> -->
<div class="is-size-5 publication-authors">
<span class="author-block">(* equal contributions, † corresponding authors) </span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block">
<sup>1</sup>
S-Lab, Nanyang Technological University
<sup>2</sup>
Shanghai Artificial Intelligence Laboratory
<br>
<sup>3</sup>
The Chinese University of Hong Kong
<sup>4</sup>
Nanjing University
</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<span class="link-block">
<a href="https://arxiv.org/abs/2311.17982" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>Paper(VBench)</span>
</a>
</span>
<span class="link-block">
<a href="https://arxiv.org/abs/2411.13503" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>Paper(VBench++)</span>
</a>
</span>
<!-- Video Link. -->
<span class="link-block">
<a href="https://www.youtube.com/watch?v=7IhCC8Qqn8Y" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-youtube"></i>
</span>
<span>Video</span>
</a>
</span>
<!-- Code Link. -->
<span class="link-block">
<a href="https://github.com/Vchitect/VBench" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fab fa-github"></i>
</span>
<span>GitHub</span>
</a>
</span>
<!-- Huggingface Demo Link. -->
<span class="link-block">
<a href="https://huggingface.co/spaces/Vchitect/VBench_Leaderboard" target="_blank"
class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<img src="assets/vbench/hf-logo.svg" style="display:block;width:330px;height:240px" />
</span>
<span>Huggingface Demo</span>
</a>
</span>
<!-- Dataset Link. -->
<!-- <span class="link-block">
<a href="" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="far fa-images"></i>
</span>
<span>Dataset</span>
</a> -->
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<!-- teaser -->
<!-- <section class="hero teaser">
<div class="container is-max-desktop">
<div class="hero-body">
<img src="assets/vbench/images/dim_results/fig_teaser_new.jpg" style="width:1000px; margin-bottom:10px" alt="Teaser."/>
<p style="margin-top: 0;">
<b>Overview of VBench.</b> We propose VBench, a comprehensive benchmark suite for video generative models. We design a comprehensive and hierarchical <b>Evaluation Dimension Suite</b> to decompose "video generation quality" into multiple well-defined dimensions to facilitate fine-grained and objective evaluation. For each dimension and each content category, we carefully design a <b>Prompt Suite</b> as test cases, and sample <b>Generated Videos</b> from a set of video generation models. For each evaluation dimension, we specifically design an <b>Evaluation Method Suite</b>, which uses carefully crafted method or designated pipeline for automatic objective evaluation. We also conduct <b>Human Preference Annotation</b> for the generated videos for each dimension, and show that VBench evaluation results are <b>well aligned with human perceptions</b>. VBench provides valuable insights and will be open-sourced.
</p>
</div>
</div>
</section> -->
<!-- teaser -->
<section class="hero teaser">
<div class="container is-max-desktop" style="width: 50%; max-width: none;">
<div class="hero-body">
<img src="assets/vbench/images/dim_results/fig_extention_teaser.jpg" style="width:100%; margin-bottom:10px" alt="Teaser."/>
<p style="margin-top: 0;">
<b>Overview of VBench.</b> We propose VBench, a comprehensive benchmark suite for video generative models. We design a comprehensive and hierarchical <b>Evaluation Dimension Suite</b> to decompose "video generation quality" into multiple well-defined dimensions to facilitate fine-grained and objective evaluation. For each dimension and each content category, we carefully design a <b>Prompt Suite</b> as test cases, and sample <b>Generated Videos</b> from a set of video generation models. For each evaluation dimension, we specifically design an <b>Evaluation Method Suite</b>, which uses carefully crafted method or designated pipeline for automatic objective evaluation. We also conduct <b>Human Preference Annotation</b> for the generated videos for each dimension, and show that VBench evaluation results are <b>well aligned with human perceptions</b>. VBench can provide valuable insights from multiple perspectives. <b>VBench++</b> supports a wide range of video generation tasks, including text-to-video and image-to-video, with an adaptive Image Suite for fair evaluation across different settings. It evaluates not only technical quality but also the trustworthiness of generative models, offering a comprehensive view of model performance. We continually incorporate more video generative models into VBench to inform the community about the evolving landscape of video generation.
</p>
</div>
</div>
</section>
<!-- Paper video. -->
<section class="hero is-light">
<div class="columns is-centered has-text-centered" style="margin-top: 0px; margin-bottom: 10px;">
<div class="column is-three-fifths">
<h2 class="title is-3">Video</h2>
<div class="publication-video">
<iframe width="560" height="315" src="https://www.youtube.com/embed/7IhCC8Qqn8Y?si=_gpWAOXWOi2OcaIJ" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen></iframe>
</div>
</div>
</div>
</section>
<!-- / Paper video. -->
<!-- Abstract. -->
<!-- <section>
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered" style="margin-top: 10px; margin-bottom: 0px;">
<div class="column is-four-fifths">
<h2 class="title is-3 is-centered">Abstract</h2>
<div class="content has-text-justified">
<p>
Video generation has witnessed significant advancements, yet evaluating these models remains a challenge. A comprehensive evaluation benchmark for video generation is indispensable for two reasons: 1) Existing metrics do not fully align with human perceptions; 2) An ideal evaluation system should provide insights to inform future developments of video generation. To this end, we present <b>VBench</b>, a comprehensive benchmark suite that dissects "video generation quality" into specific, hierarchical, and disentangled dimensions, each with tailored prompts and evaluation methods. VBench has three appealing properties: <b>1) Comprehensive Dimensions</b>: VBench comprises 16 dimensions in video generation (e.g., subject identity inconsistency, motion smoothness, temporal flickering, and spatial relationship, etc). The evaluation metrics with fine-grained levels reveal individual models' strengths and weaknesses. <b>2) Human Alignment</b>: We also provide a dataset of human preference annotations to validate our benchmarks' alignment with human perception, for each evaluation dimension respectively. <b>3) Valuable Insights</b>: We look into current models' ability across various evaluation dimensions, and various content types. We also investigate the gaps between video and image generation models. We will open-source VBench, including all prompts, evaluation methods, generated videos, and human preference annotations, and also include more video generation models in VBench to drive forward the field of video generation.
</p>
</div>
</div>
</div>
</section> -->
<!--/ Abstract. -->
<!-- Abstract. -->
<section>
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered" style="margin-top: 10px; margin-bottom: 0px;">
<div class="column is-four-fifths">
<h2 class="title is-3 is-centered">Abstract</h2>
<div class="content has-text-justified">
<p>
Video generation has witnessed significant advancements, yet evaluating these models remains a challenge. A comprehensive evaluation benchmark for video generation is indispensable for two reasons: 1) Existing metrics do not fully align with human perceptions; 2) An ideal evaluation system should provide insights to inform future developments of video generation. To this end, we present <b>VBench</b>, a comprehensive benchmark suite that dissects "video generation quality" into specific, hierarchical, and disentangled dimensions, each with tailored prompts and evaluation methods. VBench has three appealing properties: <b>1) Comprehensive Dimensions</b>: VBench comprises 16 dimensions in video generation (e.g., subject identity inconsistency, motion smoothness, temporal flickering, and spatial relationship, etc). The evaluation metrics with fine-grained levels reveal individual models' strengths and weaknesses. <b>2) Human Alignment</b>: We also provide a dataset of human preference annotations to validate our benchmarks' alignment with human perception, for each evaluation dimension respectively. <b>3) Valuable Insights</b>: We look into current models' ability across various evaluation dimensions, and various content types. We also investigate the gaps between video and image generation models. <b>4) Versatile Benchmarking</b>: VBench++ is designed to evaluate a wide range of video generation tasks, including text-to-video and image-to-video. We introduce a high-quality Image Suite with an adaptive aspect ratio to enable fair evaluations across different image-to-video generation settings. Beyond assessing technical quality, VBench++ evaluates the trustworthiness of generative models, providing a more holistic view of model performance. <b>5) Full Open-Sourcing</b>: We fully open-source VBench, including all prompts, evaluation methods, generated videos, and human preference annotations. We also continually add new video generation models to the VBench leaderboard to drive forward the field of video generation.
</p>
</div>
</div>
</div>
</section>
<!--/ Abstract. -->
<!-- Radar_big -->
<!-- <section class="section" style="margin-top:-50px; margin-bottom:50px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">VBench Evaluation Results of Video Generative Models</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/fig_paper_radar_big_new.jpg" style="width:600px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
We visualize the evaluation results of four video generation models in 16 VBench dimensions. We normalize the results per dimension for clearer comparisons.
</p>
<p class="is-size-7" style="margin-top: 0;">
The values have been normalized for better readability of the chart. The normalization process involves scaling each set of performance values to a common scale between 0.3 and 0.8. The formula used for normalization is: (value - min_value) / (max_value - min_value).
</p>
</div>
</div>
</section> -->
<!-- Radar_Open Source -->
<section class="section" style="margin-top:-50px; margin-bottom:-100px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">VBench Evaluation Results of Video Generative Models</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/fig_radar_open_source.jpg" style="width:600px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
We visualize the evaluation results of various publicly available video generation models across 16 VBench dimensions. We normalize the results per dimension for clearer comparisons.
</p>
<p class="is-size-7" style="margin-top: 0;">
The values have been normalized for better readability of the chart. The normalization process involves scaling each set of performance values to a common scale between 0.3 and 0.8. The formula used for normalization is: (value - min_value) / (max_value - min_value).
</div>
</div>
</section>
<!-- Radar_Close Source -->
<section class="section" style="margin-top:-150px; margin-bottom:50px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">VBench Evaluation Results of Gen-2 and Pika</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/fig_radar_close_source.jpg" style="width:600px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
We visualize the VBench evaluation results of Gen-2 and Pika. To enhance clarity, we include the results of VideoCrafter-1.0 and Show-1 as references.
</p>
<p class="is-size-7" style="margin-top: 0;">
The values have been normalized for better readability of the chart. The normalization process involves scaling each set of performance values to a common scale between 0.3 and 0.8. The formula used for normalization is: (value - min_value) / (max_value - min_value).
</div>
</div>
</section>
<!-- Radar_I2V_wo_SVD -->
<section class="section" style="margin-top:-150px; margin-bottom:50px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">VBench++ Evaluation Results of Image-to-Video Models</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/fig_extension_i2v_wo_svd.jpg" style="width:600px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
We visualize the evaluation results of six image-to-video generation models in VBench++'s I2V dimensions.
</p>
<p class="is-size-7" style="margin-top: 0;">
The values have been normalized for better readability of the chart. The normalization process involves scaling each set of performance values to a common scale between 0.3 and 0.8. The formula used for normalization is: (value - min_value) / (max_value - min_value).
</div>
</div>
</section>
<!-- Radar_trustworthiness_with_sd -->
<section class="section" style="margin-top:-150px; margin-bottom:50px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">VBench++ Results of the Trustworthiness Dimension</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/radar_trustworthiness_t2v_t2i_4dimensions.jpg" style="width:600px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
We visualize the trustworthiness evaluation results of visual generative models and image generation models.
</p>
<p class="is-size-7" style="margin-top: 0;">
The values have been normalized for better readability of the chart. The normalization process involves scaling each set of performance values to a common scale between 0.3 and 0.8. The formula used for normalization is: (value - min_value) / (max_value - min_value).
</div>
</div>
</section>
<!-- LeaderBoard -->
<section class="hero is-light" style="margin-top:-100px; margin-bottom:-50px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">Leaderboard</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<iframe
src="https://vchitect-vbench-leaderboard.hf.space"
frameborder="0"
width="1400"
height="700"
></iframe>
</div>
</div>
</div>
</div>
</section>
<!-- <section class="section" style="margin-top:-50px; margin-bottom:-70px;">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="section-title">
<h2 class="title is-3">VBench Evaluation Results per Dimension</h2>
</div>
<div class="table-container">
<table class="table" id="RawTable">
<thead>
<tr>
<th scope="col" style="text-align: center;">Model</th>
<th scope="col" style="text-align: center;">Subject Consistency</th>
<th scope="col" style="text-align: center;">Background Consistency</th>
<th scope="col" style="text-align: center;">Temporal Flickering</th>
<th scope="col" style="text-align: center;">Motion Smoothness</th>
<th scope="col" style="text-align: center;">Dynamic Degree</th>
<th scope="col" style="text-align: center;">Aesthetic Quality</th>
<th scope="col" style="text-align: center;">Imaging Quality</th>
<th scope="col" style="text-align: center;">Object Class</th>
</tr>
</thead>
<tbody>
<tr>
<td style="text-align: center; font-weight: bold;">LaVie</td>
<td style="text-align: center;">0.9141</td>
<td style="text-align: center; font-weight: bold;">0.9747</td>
<td style="text-align: center; font-weight: bold;">0.9830</td>
<td style="text-align: center;">0.9638</td>
<td style="text-align: center;">0.4972</td>
<td style="text-align: center; font-weight: bold;">0.5494</td>
<td style="text-align: center; font-weight: bold;">0.6190</td>
<td style="text-align: center; font-weight: bold;">0.9182</td>
</tr>
<tr>
<td style="text-align: center; font-weight: bold;">ModelScope</td>
<td style="text-align: center;">0.8987</td>
<td style="text-align: center;">0.9529</td>
<td style="text-align: center;">0.9828</td>
<td style="text-align: center;">0.9579</td>
<td style="text-align: center;">0.6639</td>
<td style="text-align: center;">0.5206</td>
<td style="text-align: center;">0.5857</td>
<td style="text-align: center;">0.8225</td>
</tr>
<tr>
<td scope="col" style="text-align: center;"><b>VideoCrafter</b></td>
<td scope="col" style="text-align: center;">0.8624</td>
<td scope="col" style="text-align: center;">0.9288</td>
<td scope="col" style="text-align: center;">0.9760</td>
<td scope="col" style="text-align: center;">0.9179</td>
<td scope="col" style="text-align: center;"><b>0.8972</b></td>
<td scope="col" style="text-align: center;">0.4441</td>
<td scope="col" style="text-align: center;">0.5722</td>
<td scope="col" style="text-align: center;">0.8734</td>
</tr>
<tr>
<td scope="col" style="text-align: center;"><b>CogVideo</b></td>
<td scope="col" style="text-align: center;"><b>0.9219</b></td>
<td scope="col" style="text-align: center;">0.9542</td>
<td scope="col" style="text-align: center;">0.9764</td>
<td scope="col" style="text-align: center;"><b>0.9647</b></td>
<td scope="col" style="text-align: center;">0.4222</td>
<td scope="col" style="text-align: center;">0.3818</td>
<td scope="col" style="text-align: center;">0.4103</td>
<td scope="col" style="text-align: center;">0.7340</td>
</tr>
</tbody>
</table>
</div>
<hr style="margin: 20px 0;">
<div class="table-container">
<table class="table" id="RawTable">
<thead>
<tr>
<th scope="col" style="text-align: center;">Model</th>
<th scope="col" style="text-align: center;">Multiple Objects</th>
<th scope="col" style="text-align: center;">Human Action</th>
<th scope="col" style="text-align: center;">Color</th>
<th scope="col" style="text-align: center;">Spatial Relationship</th>
<th scope="col" style="text-align: center;">Scene</th>
<th scope="col" style="text-align: center;">Appearance Style</th>
<th scope="col" style="text-align: center;">Temporal Style</th>
<th scope="col" style="text-align: center;">Overall Consistency</th>
</tr>
</thead>
<tbody>
<tr>
<td scope="col" style="text-align: center;"><b>LaVie</b></td>
<td scope="col" style="text-align: center;">0.3332</td>
<td scope="col" style="text-align: center;"><b>0.9680</b></td>
<td scope="col" style="text-align: center;"><b>0.8639</b></td>
<td scope="col" style="text-align: center;">0.3409</td>
<td scope="col" style="text-align: center;"><b>0.5269</b></td>
<td scope="col" style="text-align: center;"><b>0.2356</b></td>
<td scope="col" style="text-align: center;"><b>0.2593</b></td>
<td scope="col" style="text-align: center;"><b>0.2641</b></td>
</tr>
<tr>
<td scope="col" style="text-align: center;"><b>ModelScope</b></td>
<td scope="col" style="text-align: center;"><b>0.3898</b></td>
<td scope="col" style="text-align: center;">0.9240</td>
<td scope="col" style="text-align: center;">0.8172</td>
<td scope="col" style="text-align: center;">0.3368</td>
<td scope="col" style="text-align: center;">0.3926</td>
<td scope="col" style="text-align: center;">0.2339</td>
<td scope="col" style="text-align: center;">0.2537</td>
<td scope="col" style="text-align: center;">0.2567</td>
</tr>
<tr>
<td scope="col" style="text-align: center;"><b>VideoCrafter</b></td>
<td scope="col" style="text-align: center;">0.2593</td>
<td scope="col" style="text-align: center;">0.9300</td>
<td scope="col" style="text-align: center;">0.7884</td>
<td scope="col" style="text-align: center;"><b>0.3674</b></td>
<td scope="col" style="text-align: center;">0.4336</td>
<td scope="col" style="text-align: center;">0.2157</td>
<td scope="col" style="text-align: center;">0.2542</td>
<td scope="col" style="text-align: center;">0.2521</td>
</tr>
<tr>
<td scope="col" style="text-align: center;"><b>CogVideo</b></td>
<td scope="col" style="text-align: center;">0.1811</td>
<td scope="col" style="text-align: center;">0.7820</td>
<td scope="col" style="text-align: center;">0.7957</td>
<td scope="col" style="text-align: center;">0.1824</td>
<td scope="col" style="text-align: center;">0.2824</td>
<td scope="col" style="text-align: center;">0.2201</td>
<td scope="col" style="text-align: center;">0.0780</td>
<td scope="col" style="text-align: center;">0.0770</td>
</tr>
</tbody>
</table>
</div>
</div>
</div>
</section> -->
<section class="section" style="margin-top:0px; margin-bottom:-140px;">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="section-title">
<h2 class="title is-3">Video Quality Dimensions</h2>
<h2 class="subtitle has-text-centered">Display of Evaluation Examples in Different Dimensions.<br>(the larger score denotes a better performance)</h2>
</div>
<!-- Single Line -->
<div id="results-carousel-face" class="carousel results-carousel">
<div class="item item-puppet">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/flicker_and_subject.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item item-round_bird">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/backdround_and_motion.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item item-puppet">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/dynamic_and_aesthetic.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item item-round_bird">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/technical_and_flicker.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
</div>
</section>
<section class="section" style="margin-top:-100px; margin-bottom:-60px;">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="section-title">
<h2 class="title is-3">Video-Condition Consistency Dimensions</h2>
<h2 class="subtitle has-text-centered">Display of Evaluation Examples in Different Dimensions.<br>(the larger score denotes a better performance)</h2>
</div>
<!-- Single Line -->
<div id="results-carousel-face" class="carousel results-carousel">
<div class="item item-puppet">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/object_and_multiple.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item item-round_bird">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/action_and_color.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item item-puppet">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/relation_and_scene.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item item-round_bird">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/appearance_and_temporal_style.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item item-puppet">
<div class="carousel-content">
<video id="dimensions" autoplay="" muted="" loop="" playsinline="" height="100%">
<source src="assets/vbench/images/dim_results/ppt_mp4_v3/overall_and_object.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
</div>
</section>
<!-- prompt_suite_stats -->
<section class="section" style="margin-top:-100px; margin-bottom:-100px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">Prompt Suite Statistics</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/prompt_suite_stats.jpg" style="width:1000px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
The two graphs provide an overview of our prompt suites. <b>Left: </b>the word cloud to visualize word distribution of our prompt suites. <b>Right: </b>the
number of prompts across different evaluation dimensions and different content categories.
</p>
</div>
</div>
</section>
<!-- Radar_sd -->
<section class="section" style="margin-top:-100px; margin-bottom:-100px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">T2V vs. T2I</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/fig_paper_radar_sd_new.jpg" style="width:700px; margin-top:10px;margin-bottom:10px;;margin-left:40px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
We use VBench to evaluate other models and baselines for further comparative analysis of T2V models, such as text-to-image (T2I) generation models
</p>
<p class="is-size-7" style="margin-top: 0;">
The values have been normalized for better readability of the chart. The normalization process involves scaling each set of performance values to a common scale between 0.3 and 0.8. The formula used for normalization is: (value - min_value) / (max_value - min_value).
</div>
</div>
</section>
<!-- Radar_per_Class -->
<section class="section" style="margin-top:-100px; margin-bottom:-50px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">VBench Results across Eight Content Categories</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/fig_paper_radar_per_class_new.jpg" style="width:1500px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<p style="margin-top: 0;">
For each chart, we plot the VBench evaluation results across eight different content categories, benchmarked by our <i>Prompt Suite per Category</i>.
</p>
<p class="is-size-7" style="margin-top: 0;">
The results are linearly normalized between 0 and 1 for better visibility across categories.
</div>
</div>
</section>
<!-- image suite crop pipe -->
<section class="section" style="margin-top:-100px; margin-bottom:-10px;">
<div class="container is-max-desktop" style="width: 33%; max-width: none;">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">Image Suite Cropping Pipeline of VBench++</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div style="margin-top:20px; margin-bottom: 10px;">
<img src="assets/vbench/images/dim_results/fig_image_crop_pipeline_vertical.jpg" style="width: 100%;" alt="Image 1">
<p style="font-size: 18px; font-weight: bold;">(a) Cropping Pipeline for Portrait Images.</p>
</div>
<!-- 第二张图片及其标注 -->
<div>
<img src="assets/vbench/images/dim_results/fig_image_crop_pipeline_horizontal.jpg" style="width: 100%;" alt="Image 2">
<p style="font-size: 18px; font-weight: bold;">(b) Cropping Pipeline for Landscape Images.</p>
</div>
</div>
</div>
<p style="margin-top: 0;">
<b>Image Suite Pipeline for Adaptive Aspect Ratio Cropping.</b> We provide a pipeline that crops images to various aspect ratios while preserving key content. <b>(a) Portrait Images.</b> If the original image's width is less than its height, it is first cropped to a 1:1 ratio (red bounding box). followed by a second crop to a 16:9 aspect ratio (yellow bounding box). Additional crops interpolate between the 1:1 red box and the 16:9 yellow box to produce other common ratios (1:1, 7:4, 8:5, 16:9). <b>(b) Landscape Images.</b> If the original image's width is greater than its height, we first crop the image to a 16:9 aspect ratio (red bounding box), and further crop the 16:9 image to a 1:1 aspect ratio (yellow bounding box). We then perform additional crops between the 16:9 red box and 1:1 yellow box to obtain the common aspect ratios (1:1, 7:4, 8:5, 16:9).
</p>
</div>
</div>
</section>
<!-- image suite stats -->
<section class="section" style="margin-top:-100px; margin-bottom:-100px;">
<div class="container is-max-desktop">
<div class="hero-body">
<div class="section-title">
<h2 class="title is-3 is-centered">Content Distribution of Image Suite</h2>
</div>
<div class="columns is-centered has-text-centered">
<div class="column">
<div class="publication-img">
<img id="architecture" src="assets/vbench/images/dim_results/pie_chart_image_suite.png" style="width:1000px; margin-top:10px;margin-bottom:10px;"/>
</div>
</div>
</div>
<!-- <p style="margin-top: 0; margin-bottom: 30px;">
Our image suite encompasses a wide variety of content to ensure a comprehensive evaluation.
</p> -->
</div>
</div>
</section>
<!-- BibTeX -->
<section class="hero is-light is-small" id="BibTeX" >
<div class="container is-max-desktop content" style="margin-top: 40px; margin-bottom: 20px;">
<h2 class="title">BibTeX</h2>
<p>If you find our work useful, please consider citing our paper:</p>
<pre><code>@InProceedings{huang2023vbench,
title={{VBench}: Comprehensive Benchmark Suite for Video Generative Models},
author={Huang, Ziqi and He, Yinan and Yu, Jiashuo and Zhang, Fan and Si, Chenyang and Jiang, Yuming and Zhang, Yuanhan and Wu, Tianxing and Jin, Qingyang and Chanpaisit, Nattapol and Wang, Yaohui and Chen, Xinyuan and Wang, Limin and Lin, Dahua and Qiao, Yu and Liu, Ziwei},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year={2024}
}
@article{huang2024vbench++,
title={VBench++: Comprehensive and Versatile Benchmark Suite for Video Generative Models},
author={Huang, Ziqi and Zhang, Fan and Xu, Xiaojie and He, Yinan and Yu, Jiashuo and Dong, Ziyue and Ma, Qianli and Chanpaisit, Nattapol and Si, Chenyang and Jiang, Yuming and Wang, Yaohui and Chen, Xinyuan and Chen, Ying-Cong and Wang, Limin and Lin, Dahua and Qiao, Yu and Liu, Ziwei},
journal={arXiv preprint arXiv:2411.13503},
year={2024}
}</code></pre>
</div>
</section>
<!-- footer -->
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link"
href="https://vchitect.github.io/VBench-project/">
<i class="fas fa-file-pdf"></i>
</a>
<a class="icon-link" href="https://github.com/Vchitect/VBench" class="external-link" disabled>
<i class="fab fa-github"></i>
</a>
</div>
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This website is licensed under a <a rel="license"
href="http://creativecommons.org/licenses/by-sa/4.0/">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
<p>
Website source code based on the <a href="https://nerfies.github.io/"> Nerfies</a> project page. If you want to reuse their <a
href="https://github.com/nerfies/nerfies.github.io">source code</a>, please credit them appropriately.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>