SpectralNeRF: Physically Based Spectral Rendering with Neural Radiance Field

TL;DR

SpectralNeRF introduces a novel spectral rendering approach using neural radiance fields, enabling high-quality, physically-based view synthesis by decomposing scenes into spectral components and combining them for accurate RGB images.

Contribution

The paper presents SpectralNeRF, a new spectral rendering framework that enhances NeRF with spectral decomposition and attention mechanisms for improved view synthesis.

Findings

Outperforms recent NeRF-based methods on synthetic and real datasets.

Provides physically-based spectral rendering with improved scene decomposition.

Achieves higher quality and more accurate view synthesis results.

Abstract

In this paper, we propose SpectralNeRF, an end-to-end Neural Radiance Field (NeRF)-based architecture for high-quality physically based rendering from a novel spectral perspective. We modify the classical spectral rendering into two main steps, 1) the generation of a series of spectrum maps spanning different wavelengths, 2) the combination of these spectrum maps for the RGB output. Our SpectralNeRF follows these two steps through the proposed multi-layer perceptron (MLP)-based architecture (SpectralMLP) and Spectrum Attention UNet (SAUNet). Given the ray origin and the ray direction, the SpectralMLP constructs the spectral radiance field to obtain spectrum maps of novel views, which are then sent to the SAUNet to produce RGB images of white-light illumination. Applying NeRF to build up the spectral rendering is a more physically-based way from the perspective of ray-tracing. Further, the spectral radiance fields decompose difficult scenes and improve the performance of NeRF-based methods. Comprehensive experimental results demonstrate the proposed SpectralNeRF is superior to recent NeRF-based methods when synthesizing new views on synthetic and real datasets. The codes and datasets are available at https://github.com/liru0126/SpectralNeRF.