Skip to content

Commit bd70276

Browse files
Video-BenchVideo-Bench
authored
Update index.html
1 parent 181b7e8 commit bd70276

1 file changed

Lines changed: 126 additions & 132 deletions

File tree

‎index.html‎

Lines changed: 126 additions & 132 deletions
Original file line numberDiff line numberDiff line change
@@ -5,29 +5,33 @@
55
<meta name="viewport" content="width=device-width, initial-scale=1.0">
66
<meta http-equiv="X-UA-Compatible" content="ie=edge">
77
<title>Video-Bench: Human-Aligned Video Generation Benchmark</title>
8-
<link rel="stylesheet" href="styles.css"> <!-- Link to a CSS file for styling, if needed -->
98
<style>
109
body {
1110
font-family: Arial, sans-serif;
12-
line-height: 1.6;
1311
margin: 0;
1412
padding: 0;
1513
background-color: #f4f4f4;
1614
}
1715
header {
18-
background: #333;
19-
color: #fff;
20-
padding: 10px 0;
16+
background-color: #333;
17+
color: white;
2118
text-align: center;
19+
padding: 20px;
2220
}
2321
header h1 {
22+
font-size: 2.5em;
2423
margin: 0;
2524
}
25+
header h2 {
26+
font-size: 1.5em;
27+
margin: 5px 0;
28+
}
2629
section {
27-
padding: 20px;
2830
margin: 20px;
31+
padding: 20px;
2932
background-color: white;
3033
border-radius: 8px;
34+
box-shadow: 0px 0px 10px rgba(0, 0, 0, 0.1);
3135
}
3236
h2 {
3337
color: #333;
@@ -50,23 +54,13 @@
5054
a:hover {
5155
text-decoration: underline;
5256
}
53-
.video-container {
54-
text-align: center;
55-
margin-top: 20px;
56-
}
57-
.image-container {
58-
text-align: center;
59-
margin-top: 20px;
60-
}
61-
.image-container img {
62-
max-width: 100%;
63-
height: auto;
64-
border-radius: 8px;
57+
.table-container {
58+
margin: 20px 0;
6559
}
6660
table {
6761
width: 100%;
6862
border-collapse: collapse;
69-
margin: 20px 0;
63+
margin-bottom: 20px;
7064
}
7165
table, th, td {
7266
border: 1px solid #ddd;
@@ -78,146 +72,146 @@
7872
th {
7973
background-color: #f4f4f4;
8074
}
75+
.image-container {
76+
text-align: center;
77+
margin: 20px 0;
78+
}
79+
.image-container img {
80+
max-width: 100%;
81+
height: auto;
82+
border-radius: 8px;
83+
}
84+
.video-container {
85+
text-align: center;
86+
margin-top: 20px;
87+
}
88+
.video-container iframe {
89+
width: 100%;
90+
max-width: 800px;
91+
height: 450px;
92+
border-radius: 8px;
93+
}
8194
</style>
8295
</head>
8396
<body>
8497

8598
<header>
8699
<h1>Video-Bench: Human-Aligned Video Generation Benchmark</h1>
100+
<h2>CVPR 2024 Highlight</h2>
101+
<p>by Ziqi Huang, Yinan He, Jiashuo Yu, Fan Zhang, Chenyang Si, Yuming Jiang, Yuhao Wang, and others</p>
87102
</header>
88103

89104
<section>
90-
<h2>Authors</h2>
91-
<p><strong>Hui Han</strong>, <strong>Siyuan Li</strong>, <strong>Jiaqi Chen</strong>, <strong>Yiwen Yuan</strong>, <strong>Yuling Wu</strong>, <strong>Chak Tou Leong</strong>, <strong>Hanwen Du</strong>, <strong>Junchen Fu</strong>, <strong>Youhua Li</strong>, <strong>Jie Zhang</strong>, <strong>Chi Zhang</strong>, <strong>Li-jia Li</strong>, <strong>Yongxin Ni</strong></p>
92-
93-
<h2>Affiliations</h2>
94-
<ul>
95-
<li>Shanghai Jiao Tong University</li>
96-
<li>Stanford University</li>
97-
<li>Fellou AI</li>
98-
<li>Fudan University</li>
99-
<li>Carnegie Mellon University</li>
100-
<li>Hong Kong Polytechnic University</li>
101-
<li>Soochow University</li>
102-
<li>University of Glasgow</li>
103-
<li>City University of Hong Kong</li>
104-
<li>Westlake University</li>
105-
<li>LiveX AI</li>
106-
<li>National University of Singapore</li>
107-
</ul>
108-
109105
<h2>Project Overview</h2>
110106
<p>Video generation assessment is critical for ensuring generative models produce visually realistic, high-quality videos aligned with human expectations. Current video generation benchmarks are limited in aligning with human judgment. To address this, <strong>Video-Bench</strong> is introduced—a comprehensive benchmark incorporating large language models (LLMs) to evaluate video generation quality. The framework includes automated multimodal LLM evaluation, improving the alignment with human preferences. Experimental results show that Video-Bench significantly outperforms previous methods and provides more objective and accurate insights into generated video quality.</p>
111107

112108
<h2>Main Results</h2>
113109

114110
<h3>Comparison with Existing Evaluation Methods</h3>
115-
<table>
116-
<thead>
117-
<tr>
118-
<th>Model</th>
119-
<th>Video Quality</th>
120-
<th>Video-Condition Alignment</th>
121-
<th>Overall</th>
122-
</tr>
123-
</thead>
124-
<tbody>
125-
<tr>
126-
<td>Gen3</td>
127-
<td>4.66</td>
128-
<td>4.38</td>
129-
<td>1</td>
130-
</tr>
131-
<tr>
132-
<td>CogVideoX</td>
133-
<td>3.87</td>
134-
<td>4.62</td>
135-
<td>2</td>
136-
</tr>
137-
<tr>
138-
<td>VideoCrafter2</td>
139-
<td>4.08</td>
140-
<td>4.18</td>
141-
<td>3</td>
142-
</tr>
143-
<tr>
144-
<td>Kling</td>
145-
<td>4.26</td>
146-
<td>4.07</td>
147-
<td>4</td>
148-
</tr>
149-
<tr>
150-
<td>Show-1</td>
151-
<td>3.30</td>
152-
<td>4.21</td>
153-
<td>5</td>
154-
</tr>
155-
<tr>
156-
<td>LaVie</td>
157-
<td>3.00</td>
158-
<td>3.71</td>
159-
<td>6</td>
160-
</tr>
161-
</tbody>
162-
</table>
111+
<div class="table-container">
112+
<table>
113+
<thead>
114+
<tr>
115+
<th>Model</th>
116+
<th>Video Quality</th>
117+
<th>Video-Condition Alignment</th>
118+
<th>Overall Rank</th>
119+
</tr>
120+
</thead>
121+
<tbody>
122+
<tr>
123+
<td>Gen3</td>
124+
<td>4.66</td>
125+
<td>4.38</td>
126+
<td>1</td>
127+
</tr>
128+
<tr>
129+
<td>CogVideoX</td>
130+
<td>3.84</td>
131+
<td>4.62</td>
132+
<td>2</td>
133+
</tr>
134+
<tr>
135+
<td>VideoCrafter2</td>
136+
<td>4.08</td>
137+
<td>4.18</td>
138+
<td>3</td>
139+
</tr>
140+
<tr>
141+
<td>Kling</td>
142+
<td>4.26</td>
143+
<td>4.07</td>
144+
<td>4</td>
145+
</tr>
146+
<tr>
147+
<td>Show-1</td>
148+
<td>3.30</td>
149+
<td>4.21</td>
150+
<td>5</td>
151+
</tr>
152+
<tr>
153+
<td>LaVie</td>
154+
<td>3.00</td>
155+
<td>3.71</td>
156+
<td>6</td>
157+
</tr>
158+
<tr>
159+
<td>PiKa-Beta</td>
160+
<td>3.76</td>
161+
<td>2.60</td>
162+
<td>7</td>
163+
</tr>
164+
</tbody>
165+
</table>
166+
</div>
163167

164168
<h3>Human Preference Alignment Scores</h3>
165-
<table>
166-
<thead>
167-
<tr>
168-
<th>Dimension</th>
169-
<th>Video-Condition Alignment</th>
170-
<th>Video Quality</th>
171-
<th>Average Score</th>
172-
</tr>
173-
</thead>
174-
<tbody>
175-
<tr>
176-
<td>Imaging Quality</td>
177-
<td>0.733</td>
178-
<td>0.633</td>
179-
<td>0.733</td>
180-
</tr>
181-
<tr>
182-
<td>Aesthetic Quality</td>
183-
<td>0.702</td>
184-
<td>0.446</td>
185-
<td>0.702</td>
186-
</tr>
187-
<tr>
188-
<td>Motion Quality</td>
189-
<td>0.514</td>
190-
<td>0.469</td>
191-
<td>0.514</td>
192-
</tr>
193-
<tr>
194-
<td>Video-Text Consistency</td>
195-
<td>0.732</td>
196-
<td>0.611</td>
197-
<td>0.732</td>
198-
</tr>
199-
</tbody>
200-
</table>
169+
<div class="table-container">
170+
<table>
171+
<thead>
172+
<tr>
173+
<th>Entities</th>
174+
<th>Video Quality</th>
175+
<th>Video-Condition Alignment</th>
176+
<th>Average Score</th>
177+
</tr>
178+
</thead>
179+
<tbody>
180+
<tr>
181+
<td>HU - HU</td>
182+
<td>0.63</td>
183+
<td>0.47</td>
184+
<td>0.52</td>
185+
</tr>
186+
<tr>
187+
<td>HU - GPT</td>
188+
<td>0.51</td>
189+
<td>0.47</td>
190+
<td>0.41</td>
191+
</tr>
192+
<tr>
193+
<td>HU - HA</td>
194+
<td>0.61</td>
195+
<td>0.50</td>
196+
<td>0.50</td>
197+
</tr>
198+
</tbody>
199+
</table>
200+
</div>
201201

202+
<h2>Project Video Demonstration</h2>
202203
<div class="video-container">
203-
<h2>Project Video Demonstration</h2>
204-
<p>Check out the project’s demonstration video on YouTube:</p>
205-
<a href="https://youtu.be/BMvgyWbWPFg" target="_blank">
206-
<img src="https://img.youtube.com/vi/BMvgyWbWPFg/0.jpg" alt="Video Thumbnail" style="width:100%; border-radius: 8px;">
207-
</a>
204+
<iframe src="https://www.youtube.com/embed/BMvgyWbWPFg" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
208205
</div>
209206

207+
<h2>Head Image</h2>
210208
<div class="image-container">
211-
<h2>Paper's Head Image</h2>
212-
<img src="YOUR_IMAGE_URL_HERE" alt="Project Image" /> <!-- Replace this URL with the actual image URL -->
209+
<img src="dimension_v2.pdf" alt="Head Image" />
213210
</div>
214211

215212
<h2>GitHub Repository</h2>
216213
<p>For more details, visit the official repository: <a href="https://github.com/Video-Bench/Video-Bench.git" target="_blank">Video-Bench GitHub Repository</a></p>
217214

218-
<h2>Conclusion</h2>
219-
<p>The Video-Bench benchmark introduces a new paradigm for evaluating video generation models with enhanced alignment to human preferences. Through rigorous experimentation and human alignment studies, Video-Bench demonstrates its potential to become a standard evaluation tool for video generation models.</p>
220-
221215
</section>
222216

223217
</body>

0 commit comments

Comments
 (0)