UPNeRF: A Unified Framework for Monocular 3D Object Reconstruction and Pose Estimation

arXiv:2403.15705v1 Announce Type: new
Abstract: Monocular 3D reconstruction for categorical objects heavily relies on accurately perceiving each object's pose. While gradient-based optimization within a NeRF framework updates initially given poses, this paper highlights that such a scheme fails when the initial pose even moderately deviates from the true pose. Consequently, existing methods often depend on a third-party 3D object to provide an initial object pose, leading to increased complexity and generalization issues. To address these challenges, we present UPNeRF, a …

3d object 3d reconstruction abstract arxiv categorical cs.cv framework gradient highlights nerf object objects optimization paper true type updates