NeurMiPs: Neural Mixture of Planar Experts for View Synthesis
This addresses the need for efficient and flexible 3D scene modeling in computer vision, though it appears incremental as it builds on existing planar and neural radiance field methods.
The paper tackled the problem of novel view synthesis by proposing NeurMiPs, a planar-based scene representation that blends explicit mesh rendering efficiency with neural radiance field flexibility, achieving superior performance and speed compared to other 3D representations.
We present Neural Mixtures of Planar Experts (NeurMiPs), a novel planar-based scene representation for modeling geometry and appearance. NeurMiPs leverages a collection of local planar experts in 3D space as the scene representation. Each planar expert consists of the parameters of the local rectangular shape representing geometry and a neural radiance field modeling the color and opacity. We render novel views by calculating ray-plane intersections and composite output colors and densities at intersected points to the image. NeurMiPs blends the efficiency of explicit mesh rendering and flexibility of the neural radiance field. Experiments demonstrate superior performance and speed of our proposed method, compared to other 3D representations in novel view synthesis.