NeRF-Guided Unsupervised Learning of RGB-D Registration

arXiv:2405.00507v1 Announce Type: new
Abstract: This paper focuses on training a robust RGB-D registration model without ground-truth pose supervision. Existing methods usually adopt a pairwise training strategy based on differentiable rendering, which enforces the photometric and the geometric consistency between the two registered frames as supervision. However, this frame-to-frame framework suffers from poor multi-view consistency due to factors such as lighting changes, geometry occlusion and reflective materials. In this paper, we present NeRF-UR, a novel frame-to-model optimization framework for unsupervised …

abstract arxiv cs.cv differentiable framework ground-truth however nerf paper registration rendering rgb-d robust strategy supervision training truth type unsupervised unsupervised learning view