References & Citations
Computer Science > Computer Vision and Pattern Recognition
Title: Temporally-Coherent Surface Reconstruction via Metric-Consistent Atlases
(Submitted on 14 Apr 2021 (v1), last revised 23 Aug 2021 (this version, v2))
Abstract: We propose a method for the unsupervised reconstruction of a temporally-coherent sequence of surfaces from a sequence of time-evolving point clouds, yielding dense, semantically meaningful correspondences between all keyframes. We represent the reconstructed surface as an atlas, using a neural network. Using canonical correspondences defined via the atlas, we encourage the reconstruction to be as isometric as possible across frames, leading to semantically-meaningful reconstruction. Through experiments and comparisons, we empirically show that our method achieves results that exceed that state of the art in the accuracy of unsupervised correspondences and accuracy of surface reconstruction.
Submission history
From: Jan Bednařík [view email][v1] Wed, 14 Apr 2021 16:21:22 GMT (27208kb,D)
[v2] Mon, 23 Aug 2021 12:40:23 GMT (6417kb,D)
Link back to: arXiv, form interface, contact.