EVER：用于实时视图合成的精确体椭球渲染
EVER: Exact Volumetric Ellipsoid Rendering for Real-time View Synthesis
AI总结：
本文提出基于光线追踪的精确体椭球渲染方法 EVER，实现实时可微仅发射体渲染，在保持约30 FPS的同时消除3DGS伪影并提升大规模场景质量。
摘要：
我们提出了 Exact Volumetric Ellipsoid Rendering（EVER），这是一种用于实时、可微、仅发射体渲染的方法。与近期由 3D Gaussian Splatting（3DGS）提出的基于光栅化的方法不同，我们基于基元的表示允许精确的体渲染，而不是对 3D Gaussian 公告牌进行 alpha 合成。因此，与 3DGS 不同，我们的公式不会受到弹出伪影和视角依赖密度的影响，同时仍能在 NVIDIA RTX4090 上以 720p 达到约 30 FPS 的帧率。由于我们的方法建立在光线追踪之上，它能够实现散焦模糊和相机畸变（例如鱼眼相机产生的畸变）等效果，而这些效果难以通过光栅化实现。我们表明，与 3DGS 以及视图一致性渲染的后续工作相比，我们的方法更准确，混合问题更少，尤其是在 Zip-NeRF 数据集中具有挑战性的大规模场景上，它在实时技术中取得了最锐利的结果。
英文摘要：
We present Exact Volumetric Ellipsoid Rendering (EVER), a method for real-time differentiable emission-only volume rendering. Unlike recent rasterization based approach by 3D Gaussian Splatting (3DGS), our primitive based representation allows for exact volume rendering, rather than alpha compositing 3D Gaussian billboards. As such, unlike 3DGS our formulation does not suffer from popping artifacts and view dependent density, but still achieves frame rates of $\sim\!30$ FPS at 720p on an NVIDIA RTX4090. Since our approach is built upon ray tracing it enables effects such as defocus blur and camera distortion (e.g. such as from fisheye cameras), which are difficult to achieve by rasterization. We show that our method is more accurate with fewer blending issues than 3DGS and follow-up work on view-consistent rendering, especially on the challenging large-scale scenes from the Zip-NeRF dataset where it achieves sharpest results among real-time techniques.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
