Learning Generalizable Light Field Networks from Few Images

TL;DR

This paper introduces a neural light field approach for few-shot novel view synthesis that uses a 3D feature volume and achieves faster rendering speeds while maintaining competitive quality.

Contribution

It presents a new neural light field method conditioned on 3D features, enabling rapid and generalizable view synthesis from limited images.

Findings

Achieves competitive results on synthetic and real data.

Offers 100 times faster rendering than existing methods.

Effective in few-shot scenarios.

Abstract

We explore a new strategy for few-shot novel view synthesis based on a neural light field representation. Given a target camera pose, an implicit neural network maps each ray to its target pixel's color directly. The network is conditioned on local ray features generated by coarse volumetric rendering from an explicit 3D feature volume. This volume is built from the input images using a 3D ConvNet. Our method achieves competitive performances on synthetic and real MVS data with respect to state-of-the-art neural radiance field based competition, while offering a 100 times faster rendering.