DMesh++: An Efficient Differentiable Mesh for Complex Shapes

TL;DR

This paper introduces DMesh++, a novel differentiable mesh method that significantly reduces computational costs and memory usage, enabling efficient reconstruction of complex 2D and 3D shapes from various data sources.

Contribution

It presents a new differentiable mesh processing technique with improved efficiency, reducing time complexity to O(log N) and enabling complex shape reconstruction from point clouds and images.

Findings

Reduces mesh processing time complexity from O(N) to O(log N)

Requires less memory than previous methods

Successfully reconstructs complex shapes from diverse data sources

Abstract

Recent probabilistic methods for 3D triangular meshes capture diverse shapes by differentiable mesh connectivity, but face high computational costs with increased shape details. We introduce a new differentiable mesh processing method that addresses this challenge and efficiently handles meshes with intricate structures. Our method reduces time complexity from O(N) to O(log N) and requires significantly less memory than previous approaches. Building on this innovation, we present a reconstruction algorithm capable of generating complex 2D and 3D shapes from point clouds or multi-view images. Visit our project page (https://sonsang.github.io/dmesh2-project) for source code and supplementary material.