Is-NeRF: In-scattering Neural Radiance Field for Blurred Images

TL;DR

Is-NeRF introduces an in-scattering model for neural radiance fields that explicitly accounts for complex lightpaths, significantly improving the quality of scene reconstruction from motion-blurred images.

Contribution

This work presents a novel scattering-aware volume rendering pipeline and adaptive learning strategy for NeRF, enabling better handling of complex light interactions in blurry images.

Findings

Outperforms state-of-the-art methods in scene reconstruction quality.

Effectively models complex light propagation in real-world scenarios.

Accurately recovers scene details from motion-blurred images.

Abstract

Neural Radiance Fields (NeRF) has gained significant attention for its prominent implicit 3D representation and realistic novel view synthesis capabilities. Available works unexceptionally employ straight-line volume rendering, which struggles to handle sophisticated lightpath scenarios and introduces geometric ambiguities during training, particularly evident when processing motion-blurred images. To address these challenges, this work proposes a novel deblur neural radiance field, Is-NeRF, featuring explicit lightpath modeling in real-world environments. By unifying six common light propagation phenomena through an in-scattering representation, we establish a new scattering-aware volume rendering pipeline adaptable to complex lightpaths. Additionally, we introduce an adaptive learning strategy that enables autonomous determining of scattering directions and sampling intervals to capture finer object details. The proposed network jointly optimizes NeRF parameters, scattering parameters, and camera motions to recover fine-grained scene representations from blurry images. Comprehensive evaluations demonstrate that it effectively handles complex real-world scenarios, outperforming state-of-the-art approaches in generating high-fidelity images with accurate geometric details.