Invertible Neural Warp for NeRF

TL;DR

This paper introduces an invertible neural warp approach for NeRF that jointly optimizes camera pose and scene representation, using invertible neural networks to improve pose accuracy and reconstruction quality.

Contribution

It proposes a novel overparameterized, invertible neural warp model for camera pose estimation in NeRF, integrating geometry constraints for better optimization.

Findings

Outperforms existing methods in pose estimation accuracy

Achieves higher-fidelity scene reconstruction

Demonstrates robust optimization convergence

Abstract

This paper tackles the simultaneous optimization of pose and Neural Radiance Fields (NeRF). Departing from the conventional practice of using explicit global representations for camera pose, we propose a novel overparameterized representation that models camera poses as learnable rigid warp functions. We establish that modeling the rigid warps must be tightly coupled with constraints and regularization imposed. Specifically, we highlight the critical importance of enforcing invertibility when learning rigid warp functions via neural network and propose the use of an Invertible Neural Network (INN) coupled with a geometry-informed constraint for this purpose. We present results on synthetic and real-world datasets, and demonstrate that our approach outperforms existing baselines in terms of pose estimation and high-fidelity reconstruction due to enhanced optimization convergence.