Active Learning of Neural Collision Handler for Complex 3D Mesh Deformations

TL;DR

This paper introduces a deep learning-based collision detection and handling system for complex 3D mesh deformations, capable of generalizing to unseen high-dimensional meshes with improved accuracy and lower false negatives.

Contribution

It proposes a bilevel autoencoder with attention for collision detection, combined with progressive data refinement for robust handling of complex meshes.

Findings

Achieves 93.8-98.1% accuracy compared to analytic groundtruth.

Reduces false negative rate by 5.16%-25.50% over prior methods.

Increases collision handling success rate by 9.65%-58.91%.

Abstract

We present a robust learning algorithm to detect and handle collisions in 3D deforming meshes. Our collision detector is represented as a bilevel deep autoencoder with an attention mechanism that identifies colliding mesh sub-parts. We use a numerical optimization algorithm to resolve penetrations guided by the network. Our learned collision handler can resolve collisions for unseen, high-dimensional meshes with thousands of vertices. To obtain stable network performance in such large and unseen spaces, we progressively insert new collision data based on the errors in network inferences. We automatically label these data using an analytical collision detector and progressively fine-tune our detection networks. We evaluate our method for collision handling of complex, 3D meshes coming from several datasets with different shapes and topologies, including datasets corresponding to dressed and undressed human poses, cloth simulations, and human hand poses acquired using multiview capture systems. Our approach outperforms supervised learning methods and achieves $93.8-98.1\%$ accuracy compared to the groundtruth by analytic methods. Compared to prior learning methods, our approach results in a $5.16\%-25.50\%$ lower false negative rate in terms of collision checking and a $9.65\%-58.91\%$ higher success rate in collision handling.