Nonwatertight Mesh Reconstruction

TL;DR

This paper introduces a novel method for reconstructing non-watertight 3D meshes from point clouds by extending existing watertight reconstruction techniques, framing it as a semantic segmentation problem.

Contribution

It presents a new approach that adapts learning-based watertight mesh reconstruction methods to handle non-watertight meshes, a previously unexplored area.

Findings

Achieves compelling results compared to baseline techniques

Successfully extends watertight reconstruction pipeline to non-watertight meshes

Frames the problem as semantic segmentation for effective surface detection

Abstract

Reconstructing 3D non-watertight mesh from an unoriented point cloud is an unexplored area in computer vision and computer graphics. In this project, we tried to tackle this problem by extending the learning-based watertight mesh reconstruction pipeline presented in the paper 'Shape as Points'. The core of our approach is to cast the problem as a semantic segmentation problem that identifies the region in the 3D volume where the mesh surface lies and extracts the surfaces from the detected regions. Our approach achieves compelling results compared to the baseline techniques.