ExtraNeRF: Visibility-Aware View Extrapolation of Neural Radiance Fields with Diffusion Models

TL;DR

ExtraNeRF combines scene-specific NeRF modeling with diffusion-based inpainting to achieve high-quality view extrapolation and disocclusion in novel viewpoints, outperforming prior methods.

Contribution

It introduces a visibility-aware diffusion inpainting framework that enhances NeRF extrapolation by focusing on unobserved regions, with a two-stage diffusion process for sharpening results.

Findings

High-quality view extrapolation from few input views

Effective disocclusion in novel viewpoints

Significant improvements over prior view synthesis methods

Abstract

We propose ExtraNeRF, a novel method for extrapolating the range of views handled by a Neural Radiance Field (NeRF). Our main idea is to leverage NeRFs to model scene-specific, fine-grained details, while capitalizing on diffusion models to extrapolate beyond our observed data. A key ingredient is to track visibility to determine what portions of the scene have not been observed, and focus on reconstructing those regions consistently with diffusion models. Our primary contributions include a visibility-aware diffusion-based inpainting module that is fine-tuned on the input imagery, yielding an initial NeRF with moderate quality (often blurry) inpainted regions, followed by a second diffusion model trained on the input imagery to consistently enhance, notably sharpen, the inpainted imagery from the first pass. We demonstrate high-quality results, extrapolating beyond a small number of (typically six or fewer) input views, effectively outpainting the NeRF as well as inpainting newly disoccluded regions inside the original viewing volume. We compare with related work both quantitatively and qualitatively and show significant gains over prior art.