Temporally-Coherent Surface Reconstruction via Metric-Consistent Atlases

TL;DR

This paper introduces an unsupervised method for reconstructing temporally-coherent surfaces from evolving point clouds using neural network atlases, improving correspondence accuracy and surface quality.

Contribution

It presents a novel atlas-based neural network approach for unsupervised, temporally-coherent surface reconstruction with semantic correspondence across frames.

Findings

Outperforms state-of-the-art in correspondence accuracy

Achieves higher surface reconstruction quality

Demonstrates effectiveness on dynamic point cloud sequences

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.