Diff4Splat：复用视频扩散模型进行动态场景生成
Diff4Splat: Repurposing Video Diffusion Models for Dynamic Scene Generation
把视频扩散模型的生成先验直接前馈映射为可变形3D高斯场，单次前向即可从单图+相机轨迹得到动态场景，省去逐场景优化，可作为动态新视角合成与4D内容生成的免优化骨干。
原文依据 directly predicts a dynamic scene represented by a deformable 3D Gaussian field
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论我们提出 Diff4Splat，一个从单张图像生成动态场景的前馈框架。该方法将视频扩散模型强大的生成先验与从大规模 4D 数据集中学习到的几何和运动约束相结合。给定单张图像、一条相机轨迹以及可选的文本提示，模型直接预测以可变形 3D 高斯场表示的动态场景。该方式在一次前向中同时捕捉外观、几何与运动，无需测试时优化或后处理。框架的核心是一个视频潜变量 Transformer，它增强现有视频扩散模型，使其能够联合建模时空依赖并随时间预测 3D 高斯基元。在面向外观保真度、几何精度与运动一致性的目标监督下，Diff4Splat 可在 30 秒内生成高保真动态场景。我们在视频生成、新视角合成与几何提取任务上验证了 Diff4Splat 的有效性，其性能匹配或超越基于优化的动态场景合成方法，同时效率显著更高。
英文原文摘要 1191 字符
We introduce Diff4Splat, a feed-forward framework for dynamic scene generation from a single image. Our method synergizes the powerful generative priors of video diffusion models with geometric and motion constraints learned from a large-scale 4D dataset. Given a single image, a camera trajectory, and an optional text prompt, our model directly predicts a dynamic scene represented by a deformable 3D Gaussian field. This approach captures appearance, geometry, and motion in a single pass, eliminating the need for test-time optimization or post-hoc processing. At the core of our framework is a video latent transformer that enhances existing video diffusion models, enabling them to jointly model spatio-temporal dependencies and predict 3D Gaussian Primitives over time. Supervised by objectives targeting appearance fidelity, geometric accuracy, and motion consistency, Diff4Splat generates high-fidelity dynamic scenes within 30 seconds. We demonstrate the effectiveness of Diff4Splat across video generation, novel view synthesis, and geometry extraction, where it matches or surpasses optimization-based methods for dynamic scene synthesis while being significantly more efficient.