-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathindex.html
386 lines (343 loc) · 16.1 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
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>EGVD: Event-Guided Video Diffusion Model</title>
<link href="https://fonts.googleapis.com/css2?family=Source+Serif+Pro:wght@400;600;700&family=Source+Sans+Pro:wght@300;400;600&family=JetBrains+Mono:wght@400;500&display=swap" rel="stylesheet">
<style>
:root {
--primary: #0f5c81;
--secondary: #1d7874;
--bg-gradient-start: #1e5c96;
--bg-gradient-end: #1d7874;
--text-light: #f8fafc;
--text-dark: #2c3e50;
--text-medium: #34495e;
--card-bg: #ffffff;
--border: #e2e8f0;
--background: #f7f9fc;
}
* {
margin: 0;
padding: 0;
box-sizing: border-box;
}
body {
font-family: 'Source Sans Pro', -apple-system, BlinkMacSystemFont, sans-serif;
background-color: var(--background);
color: var(--text-medium);
line-height: 1.6;
font-size: 17px;
font-weight: 400;
padding: 2rem 0;
}
.container {
max-width: 1000px;
margin: 0 auto;
padding: 0 20px;
display: flex;
flex-direction: column;
align-items: center;
}
.header-card {
width: 100%;
max-width: 800px;
background: linear-gradient(135deg, var(--bg-gradient-start), var(--bg-gradient-end));
color: white;
border-radius: 8px;
box-shadow: 0 10px 25px rgba(0, 0, 0, 0.1);
overflow: hidden;
position: relative;
margin-bottom: 3rem;
padding: 3rem 2rem;
text-align: center;
}
.header-card::before {
content: "";
position: absolute;
top: 0;
left: 0;
right: 0;
bottom: 0;
background-image: radial-gradient(circle at 20% 30%, rgba(255, 255, 255, 0.1) 0%, transparent 25%),
radial-gradient(circle at 80% 70%, rgba(255, 255, 255, 0.08) 0%, transparent 25%);
pointer-events: none;
}
.title {
font-family: 'Source Serif Pro', Georgia, serif;
font-size: 2.4rem;
font-weight: 700;
line-height: 1.2;
margin-bottom: 1.2rem;
text-shadow: 0 2px 4px rgba(0, 0, 0, 0.2);
}
.subtitle {
font-size: 1.1rem;
font-weight: 400;
max-width: 700px;
margin: 0 auto 2rem;
opacity: 0.9;
}
.authors {
display: flex;
flex-wrap: wrap;
justify-content: center;
gap: 0.5rem 1rem;
margin-bottom: 1rem;
font-weight: 600;
}
.affiliations {
display: flex;
flex-wrap: wrap;
justify-content: center;
gap: 0.5rem;
margin-bottom: 2rem;
font-size: 0.9rem;
opacity: 0.9;
}
.buttons {
display: flex;
justify-content: center;
gap: 1rem;
margin-top: 2rem;
}
.btn {
background: white;
color: var(--primary);
border: none;
border-radius: 4px;
padding: 0.7rem 1.5rem;
font-size: 0.95rem;
font-weight: 600;
cursor: pointer;
transition: all 0.2s ease;
text-decoration: none;
display: inline-flex;
align-items: center;
gap: 0.5rem;
box-shadow: 0 4px 10px rgba(0, 0, 0, 0.1);
}
.btn:hover {
transform: translateY(-2px);
box-shadow: 0 6px 15px rgba(0, 0, 0, 0.15);
}
.btn svg {
width: 18px;
height: 18px;
}
.content-card {
width: 100%;
max-width: 800px;
background: white;
border-radius: 8px;
box-shadow: 0 5px 20px rgba(0, 0, 0, 0.05);
margin-bottom: 2rem;
overflow: hidden;
}
.content-image {
width: 100%;
height: auto;
display: block;
}
.image-caption {
padding: 1rem;
font-size: 0.9rem;
color: var(--text-medium);
text-align: center;
font-style: italic;
background: rgba(247, 249, 252, 0.7);
}
.section {
padding: 2.5rem;
}
.section-title {
font-family: 'Source Serif Pro', Georgia, serif;
color: var(--primary);
font-size: 1.8rem;
font-weight: 600;
margin-bottom: 1.5rem;
padding-bottom: 0.5rem;
position: relative;
}
.section-title::after {
content: "";
position: absolute;
bottom: 0;
left: 0;
width: 50px;
height: 3px;
background: var(--primary);
border-radius: 3px;
}
.section-content {
color: var(--text-dark);
font-size: 1rem;
line-height: 1.7;
text-align: justify;
}
.citation {
font-family: 'JetBrains Mono', monospace;
font-size: 0.85rem;
background: #f8fafc;
padding: 1.5rem;
border-radius: 8px;
border: 1px solid #e2e8f0;
white-space: pre-wrap;
overflow-x: auto;
margin-top: 2rem;
}
footer {
margin-top: 3rem;
text-align: center;
font-size: 0.9rem;
color: var(--text-medium);
}
footer a {
color: var(--primary);
text-decoration: none;
font-weight: 600;
}
footer a:hover {
text-decoration: underline;
}
@media (max-width: 768px) {
.title {
font-size: 1.8rem;
}
.subtitle {
font-size: 1rem;
}
.authors {
flex-direction: column;
align-items: center;
}
.buttons {
flex-direction: column;
align-items: center;
}
.btn {
width: 100%;
}
.section {
padding: 1.5rem;
}
}
</style>
</head>
<body>
<div class="container">
<header class="header-card">
<h1 class="title">EGVD: Event-Guided Video Diffusion Model for Physically Realistic Large-Motion Frame Interpolation</h1>
<p class="subtitle">A novel framework leveraging event camera data with stable video diffusion models for high-quality frame interpolation in challenging scenarios</p>
<div class="authors">
<span>Ziran Zhang<sup>1,2</sup></span>
<span>Xiaohui Li<sup>2,3</sup></span>
<span>Yihao Liu<sup>2</sup></span>
<span>Yujin Wang<sup>2</sup></span>
<span>Yueting Chen<sup>1</sup></span>
<span>Tianfan Xue<sup>4,2*</sup></span>
<span>Shi Guo<sup>2*</sup></span>
</div>
<div class="affiliations">
<span><sup>1</sup>Zhejiang University</span>
<span><sup>2</sup>Shanghai AI Laboratory</span>
<span><sup>3</sup>Shanghai Jiao Tong University</span>
<span><sup>4</sup>The Chinese University of Hong Kong</span>
<br><span><sup>*</sup>Corresponding authors</span>
</div>
<div class="buttons">
<a href="https://arxiv.org/abs/2503.20268" class="btn" target="_blank">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round">
<path d="M14 2H6a2 2 0 0 0-2 2v16a2 2 0 0 0 2 2h12a2 2 0 0 0 2-2V8z"></path>
<polyline points="14 2 14 8 20 8"></polyline>
<line x1="16" y1="13" x2="8" y2="13"></line>
<line x1="16" y1="17" x2="8" y2="17"></line>
</svg>
Paper
</a>
<a href="https://github.com/OpenImagingLab/EGVD" class="btn" target="_blank">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round">
<path d="M9 19c-5 1.5-5-2.5-7-3m14 6v-3.87a3.37 3.37 0 0 0-.94-2.61c3.14-.35 6.44-1.54 6.44-7A5.44 5.44 0 0 0 20 4.77 5.07 5.07 0 0 0 19.91 1S18.73.65 16 2.48a13.38 13.38 0 0 0-7 0C6.27.65 5.09 1 5.09 1A5.07 5.07 0 0 0 5 4.77a5.44 5.44 0 0 0-1.5 3.78c0 5.42 3.3 6.61 6.44 7A3.37 3.37 0 0 0 9 18.13V22"></path>
</svg>
Code
</a>
<a href="https://drive.google.com/file/d/14nGFibda426PAEWHNfr5_jcO1NG3QEyr/view?usp=drive_link" class="btn" target="_blank">
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round">
<polygon points="23 7 16 12 23 17 23 7"></polygon>
<rect x="1" y="5" width="15" height="14" rx="2" ry="2"></rect>
</svg>
Demo Video
</a>
</div>
</header>
<div class="content-card">
<div class="section">
<h2 class="section-title">Results</h2>
<div class="section-content">
<div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden; border-radius: 8px; margin-bottom: 1.5rem;">
<iframe style="position: absolute; top: 0; left: 0; width: 100%; height: 100%; border: 0;" src="https://drive.google.com/file/d/14nGFibda426PAEWHNfr5_jcO1NG3QEyr/preview" allowfullscreen></iframe>
</div>
<p>Our approach significantly outperforms existing methods in handling large motion and challenging lighting conditions. The video demonstrates the superiority of EGVD in generating physically realistic intermediate frames, particularly in scenarios with complex motion patterns.</p>
</div>
</div>
</div>
<div class="content-card">
<img src="./docs/teaser.png" alt="EGVD Teaser" class="content-image">
<div class="image-caption">Visual comparisons of our EGVD method against existing approaches for frame interpolation across diverse scenarios</div>
</div>
<div class="content-card">
<div class="section">
<h2 class="section-title">Abstract</h2>
<div class="section-content">
<p>Video frame interpolation (VFI) in scenarios with large motion remains challenging due to motion ambiguity between frames. While event cameras can capture high temporal resolution motion information, existing event-based VFI methods struggle with limited training data and complex motion patterns. In this paper, we introduce Event-Guided Video Diffusion Model (EGVD), a novel framework that leverages the powerful priors of pre-trained stable video diffusion models alongside the precise temporal information from event cameras. Our approach features a Multi-modal Motion Condition Generator (MMCG) that effectively integrates RGB frames and event signals to guide the diffusion process, producing physically realistic intermediate frames. We employ a selective fine-tuning strategy that preserves spatial modeling capabilities while efficiently incorporating event-guided temporal information. We incorporate input-output normalization techniques inspired by recent advances in diffusion modeling to enhance training stability across varying noise levels. To improve generalization, we construct a comprehensive dataset combining both real and simulated event data across diverse scenarios. Extensive experiments on both real and simulated datasets demonstrate that EGVD significantly outperforms existing methods in handling large motion and challenging lighting conditions, achieving substantial improvements in perceptual quality metrics (27.4% better LPIPS on Prophesee and 24.1% on BSRGB) while maintaining competitive fidelity measures.</p>
</div>
</div>
</div>
<div class="content-card">
<div class="section">
<h2 class="section-title">Method</h2>
<div class="section-content">
<img src="./docs/framework.png" alt="EGVD Framework" class="content-image" style="margin-bottom: 1rem;">
<div class="image-caption" style="margin-bottom: 1rem;">Overview of our Event-Guided Video Diffusion Model (EGVD) framework</div>
<p>Our EGVD framework introduces a novel approach that leverages both event camera data and stable video diffusion models to achieve high-quality frame interpolation. The key component of our method is the Multi-modal Motion Condition Generator (MMCG), which effectively integrates RGB frames and event signals to guide the diffusion process.</p>
</div>
</div>
</div>
<div class="content-card">
<div class="section">
<h2 class="section-title">Key Contributions</h2>
<div class="section-content">
<ul style="list-style-type: none; padding-left: 0;">
<li style="margin-bottom: 1rem; padding-left: 1.5rem; position: relative;">
<span style="position: absolute; left: 0; color: var(--primary);">•</span>
A novel Multi-Modal Motion Condition Generator (MMCG) that integrates event information into the SVD framework to improve the interpolation of large motions.
</li>
<li style="margin-bottom: 1rem; padding-left: 1.5rem; position: relative;">
<span style="position: absolute; left: 0; color: var(--primary);">•</span>
A two-stage training strategy that first trains the conditioning generator independently, followed by fine-tuning the SVD model to adapt to Event-VFI.
</li>
<li style="margin-bottom: 1rem; padding-left: 1.5rem; position: relative;">
<span style="position: absolute; left: 0; color: var(--primary);">•</span>
A diverse and comprehensive training dataset that combines real-world and synthetic event-RGB data, improving the generalization ability of our model.
</li>
<li style="margin-bottom: 1rem; padding-left: 1.5rem; position: relative;">
<span style="position: absolute; left: 0; color: var(--primary);">•</span>
Extensive experimental results demonstrate that our approach outperforms existing methods, particularly in large-motion and low-light scenarios.
</li>
</ul>
</div>
</div>
</div>
<div class="citation">@article{zhang2025egvd,
title={EGVD: Event-Guided Video Diffusion Model for Physically Realistic Large-Motion Frame Interpolation},
author={Zhang, Ziran and Li, Xiaohui and Liu, Yihao and Wang, Yujin and Chen, Yueting and Xue, Tianfan and Guo, Shi},
journal={arXiv preprint arXiv:2503.20268},
year={2025}
}</div>
<footer>
<p>© 2025 <a href="https://github.com/OpenImagingLab/EGVD" target="_blank">OpenImagingLab</a>. All rights reserved.</p>
</footer>
</div>
</body>
</html>