DiffusioNeRF: Regularizing Neural Radiance Fields with Denoising Diffusion Models

TL;DR

DiffusioNeRF introduces a novel regularization approach for Neural Radiance Fields using a denoising diffusion model trained on RGBD patches, improving scene geometry and color accuracy, especially with limited input views.

Contribution

This paper presents a new method that integrates a denoising diffusion model as a prior to regularize NeRF training, enhancing reconstruction quality and generalization.

Findings

Improved geometry reconstruction on LLFF dataset.

Enhanced generalization to novel views.

Better reconstruction quality on DTU dataset.

Abstract

Under good conditions, Neural Radiance Fields (NeRFs) have shown impressive results on novel view synthesis tasks. NeRFs learn a scene's color and density fields by minimizing the photometric discrepancy between training views and differentiable renderings of the scene. Once trained from a sufficient set of views, NeRFs can generate novel views from arbitrary camera positions. However, the scene geometry and color fields are severely under-constrained, which can lead to artifacts, especially when trained with few input views. To alleviate this problem we learn a prior over scene geometry and color, using a denoising diffusion model (DDM). Our DDM is trained on RGBD patches of the synthetic Hypersim dataset and can be used to predict the gradient of the logarithm of a joint probability distribution of color and depth patches. We show that, these gradients of logarithms of RGBD patch priors serve to regularize geometry and color of a scene. During NeRF training, random RGBD patches are rendered and the estimated gradient of the log-likelihood is backpropagated to the color and density fields. Evaluations on LLFF, the most relevant dataset, show that our learned prior achieves improved quality in the reconstructed geometry and improved generalization to novel views. Evaluations on DTU show improved reconstruction quality among NeRF methods.