Lighting up NeRF via Unsupervised Decomposition and Enhancement
This addresses the challenge of view inconsistency in low-light 3D scene reconstruction for applications like photography and VR, though it is an incremental improvement combining decomposition with NeRF.
The paper tackles the problem of synthesizing high-quality novel views from low-light images using Neural Radiance Fields (NeRF), which typically fail due to noise and distortion; the proposed LLNeRF method enhances scene representation directly from sRGB images, outperforming existing methods in experiments.
Neural Radiance Field (NeRF) is a promising approach for synthesizing novel views, given a set of images and the corresponding camera poses of a scene. However, images photographed from a low-light scene can hardly be used to train a NeRF model to produce high-quality results, due to their low pixel intensities, heavy noise, and color distortion. Combining existing low-light image enhancement methods with NeRF methods also does not work well due to the view inconsistency caused by the individual 2D enhancement process. In this paper, we propose a novel approach, called Low-Light NeRF (or LLNeRF), to enhance the scene representation and synthesize normal-light novel views directly from sRGB low-light images in an unsupervised manner. The core of our approach is a decomposition of radiance field learning, which allows us to enhance the illumination, reduce noise and correct the distorted colors jointly with the NeRF optimization process. Our method is able to produce novel view images with proper lighting and vivid colors and details, given a collection of camera-finished low dynamic range (8-bits/channel) images from a low-light scene. Experiments demonstrate that our method outperforms existing low-light enhancement methods and NeRF methods.