Neural NeRF Compression

TL;DR

This paper introduces a neural compression method for grid-based Neural Radiance Fields (NeRFs) that significantly reduces storage overhead while maintaining high reconstruction quality, using an encoder-free, end-to-end optimized approach with importance weighting and sparse entropy modeling.

Contribution

It proposes a novel neural compression technique for NeRF feature grids that is encoder-free and end-to-end optimized, addressing storage issues in scene-specific NeRF models.

Findings

Outperforms existing NeRF compression methods in efficiency

Achieves higher reconstruction quality with lower storage

Effective handling of spatial inhomogeneity in feature grids

Abstract

Neural Radiance Fields (NeRFs) have emerged as powerful tools for capturing detailed 3D scenes through continuous volumetric representations. Recent NeRFs utilize feature grids to improve rendering quality and speed; however, these representations introduce significant storage overhead. This paper presents a novel method for efficiently compressing a grid-based NeRF model, addressing the storage overhead concern. Our approach is based on the non-linear transform coding paradigm, employing neural compression for compressing the model's feature grids. Due to the lack of training data involving many i.i.d scenes, we design an encoder-free, end-to-end optimized approach for individual scenes, using lightweight decoders. To leverage the spatial inhomogeneity of the latent feature grids, we introduce an importance-weighted rate-distortion objective and a sparse entropy model employing a masking mechanism. Our experimental results validate that our proposed method surpasses existing works in terms of grid-based NeRF compression efficacy and reconstruction quality.