English(EN)AniGS: Bridging Rendering and Diffusion Prior for 3D Scene Animation
新的AI方法增强三维渲染、重建和动画 · 已追踪4个来源
作者PulseAugur 编辑部·[7 个来源]·
研究人员正在开发新的方法，通过整合扩散模型和先进的优化技术来增强三维渲染和重建。一种名为特征引导扩散演化（FIDE）的方法，使用Vision Transformers提取的视觉特征来指导扩散模型和CMA演化策略进行不可微逆渲染，其性能优于传统的基于梯度的方法。另一项开发是FutureSurf，它提出了一个用于动态表面重建的基准和数据集，并指出当前方法在预测观察时间窗口之外的未来几何形状方面存在困难。此外，Texture++提出了一种用于在UV空间中放大三维资产纹理的区域感知扩散模型，而AniGS则利用扩散先验来动画化由高斯溅射表示的大型三维场景，为静态重建添加细微的动态效果。
AI
Numerous 3D assets are discarded due to low texture resolution, while current super-resolution models ignore texture maps and focus on natural images. An efficient and generalizable texture super-resolution model can revitalize a large corpus of aging yet valuable assets across i…
Dynamic-scene reconstruction is almost always evaluated inside the observed time window, yet deployment settings such as AR overlays, robot interaction, and anticipatory planning need the future surface: the geometry at times beyond those captured. No standard benchmark measures …
arXiv cs.LG
TIER_1English(EN)·Andrei-Timotei Ardelean, Michael Fischer, Tim Weyrich, Tom\'a\v{s} Iser·
arXiv:2607.17411v1 Announce Type: cross Abstract: Inverse rendering is traditionally solved via differentiable renderers and gradient descent, which requires substantial problem-specific engineering and is prone to getting stuck in local minima due to ambiguities. Derivative-free…
Novel view rendering of large and complex reconstructed scenes is becoming increasingly photorealistic. However, most reconstructions remain static and lack the ambient motion that makes environments immersive. We present AniGS, a method for scene-level animation of 3D Gaussian S…
arXiv:2607.21504v1 Announce Type: new Abstract: Numerous 3D assets are discarded due to low texture resolution, while current super-resolution models ignore texture maps and focus on natural images. An efficient and generalizable texture super-resolution model can revitalize a la…
arXiv:2607.21471v1 Announce Type: new Abstract: Dynamic-scene reconstruction is almost always evaluated inside the observed time window, yet deployment settings such as AR overlays, robot interaction, and anticipatory planning need the future surface: the geometry at times beyond…
arXiv cs.CV
TIER_1English(EN)·Yen-Chi Cheng, Chen Gao, Chuhan Chen, Tuotuo Li, Rajvi Shah, Ayush Saraf, Changil Kim, Liangyan Gui, Alexander Schwing, Johannes Kopf, Hung-Yu Tseng·
arXiv:2607.18539v1 Announce Type: new Abstract: Novel view rendering of large and complex reconstructed scenes is becoming increasingly photorealistic. However, most reconstructions remain static and lack the ambient motion that makes environments immersive. We present AniGS, a m…
