index.html

<!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>