MVSNeRF: Fast Generalizable Radiance Field Reconstruction from Multi-View Stereo

TLDR

We present MVSNeRF, a neural rendering approach that efficiently reconstructs neural radiance fields from only three nearby input views via fast network inference. The method uses plane‑swept cost volumes for geometry‑aware reasoning and physically based volume rendering, trained on real objects in the DTU dataset and tested on three datasets to assess generalizability. MVSNeRF generalizes across scenes, producing realistic view synthesis from three images and outperforming concurrent generalizable radiance field methods, while enabling fast per‑scene fine‑tuning with higher rendering quality and less optimization time than NeRF when dense images are available.

Abstract

We present MVSNeRF, a novel neural rendering approach that can efficiently reconstruct neural radiance fields for view synthesis. Unlike prior works on neural radiance fields that consider per-scene optimization on densely captured images, we propose a generic deep neural network that can reconstruct radiance fields from only three nearby input views via fast network inference. Our approach leverages plane-swept cost volumes (widely used in multi-view stereo) for geometry-aware scene reasoning, and combines this with physically based volume rendering for neural radiance field reconstruction. We train our network on real objects in the DTU dataset, and test it on three different datasets to evaluate its effectiveness and generalizability. Our approach can generalize across scenes (even indoor scenes, completely different from our training scenes of objects) and generate realistic view synthesis results using only three input images, significantly outperforming concurrent works on generalizable radiance field reconstruction. Moreover, if dense images are captured, our estimated radiance field representation can be easily fine-tuned; this leads to fast per-scene reconstruction with higher rendering quality and substantially less optimization time than NeRF.