NeurMiPs: Neural Mixture of Planar Experts for View Synthesis

04/28/2022
by   Zhi-Hao Lin, et al.
14

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.

READ FULL TEXT

page 14

page 15

page 16

page 17

page 18

page 19

page 20

page 21

research
01/13/2022

Stereo Magnification with Multi-Layer Images

Representing scenes with multiple semi-transparent colored layers has be...
research
06/10/2022

Generalizable Neural Radiance Fields for Novel View Synthesis with Transformer

We propose a Transformer-based NeRF (TransNeRF) to learn a generic neura...
research
08/22/2023

Enhancing NeRF akin to Enhancing LLMs: Generalizable NeRF Transformer with Mixture-of-View-Experts

Cross-scene generalizable NeRF models, which can directly synthesize nov...
research
02/21/2022

Geometry-Aware Planar Embedding of Treelike Structures

The growing complexity of spatial and structural information in 3D data ...
research
02/17/2023

MixNeRF: Modeling a Ray with Mixture Density for Novel View Synthesis from Sparse Inputs

Neural Radiance Field (NeRF) has broken new ground in the novel view syn...
research
07/20/2023

Urban Radiance Field Representation with Deformable Neural Mesh Primitives

Neural Radiance Fields (NeRFs) have achieved great success in the past f...
research
01/25/2023

Ultra-NeRF: Neural Radiance Fields for Ultrasound Imaging

We present a physics-enhanced implicit neural representation (INR) for u...

Please sign up or login with your details

Forgot password? Click here to reset