Differentiable Voxelization and Mesh Morphing

TL;DR

This paper introduces a differentiable voxelization method for 3D meshes using winding numbers and solid angles, enabling gradient computation and fast processing, with applications in mesh morphing.

Contribution

It presents a novel differentiable voxelization technique that allows for efficient mesh deformation via neural networks, improving accuracy and speed.

Findings

Achieves state-of-the-art accuracy on ShapeNet dataset

Supports GPU acceleration for fast processing

Enables gradient-based mesh morphing

Abstract

In this paper, we propose the differentiable voxelization of 3D meshes via the winding number and solid angles. The proposed approach achieves fast, flexible, and accurate voxelization of 3D meshes, admitting the computation of gradients with respect to the input mesh and GPU acceleration. We further demonstrate the application of the proposed voxelization in mesh morphing, where the voxelized mesh is deformed by a neural network. The proposed method is evaluated on the ShapeNet dataset and achieves state-of-the-art performance in terms of both accuracy and efficiency.