Planar Surface Reconstruction from Sparse Views
This addresses the problem of reconstructing indoor environments for applications like robotics or AR/VR, though it appears incremental by focusing on planes as a specific structure.
The paper tackles planar surface reconstruction of indoor scenes from two views with unknown camera poses, achieving state-of-the-art results on challenging Matterport3D scenes.
The paper studies planar surface reconstruction of indoor scenes from two views with unknown camera poses. While prior approaches have successfully created object-centric reconstructions of many scenes, they fail to exploit other structures, such as planes, which are typically the dominant components of indoor scenes. In this paper, we reconstruct planar surfaces from multiple views, while jointly estimating camera pose. Our experiments demonstrate that our method is able to advance the state of the art of reconstruction from sparse views, on challenging scenes from Matterport3D. Project site: https://jinlinyi.github.io/SparsePlanes/