Training-free zero-shot 3D symmetry detection with visual features back-projected to geometry

TL;DR

This paper introduces a training-free, zero-shot method for 3D symmetry detection using visual features from foundation models, which are backprojected onto geometry to identify symmetry planes effectively.

Contribution

It proposes a novel approach leveraging foundation vision models for 3D symmetry detection without training, outperforming traditional and learning-based methods.

Findings

Outperforms traditional geometric methods on ShapeNet

Effective zero-shot symmetry detection without training data

Utilizes features from foundation vision models for 3D geometry

Abstract

We present a simple yet effective training-free approach for zero-shot 3D symmetry detection that leverages visual features from foundation vision models such as DINOv2. Our method extracts features from rendered views of 3D objects and backprojects them onto the original geometry. We demonstrate the symmetric invariance of these features and use them to identify reflection-symmetry planes through a proposed algorithm. Experiments on a subset of ShapeNet demonstrate that our approach outperforms both traditional geometric methods and learning-based approaches without requiring any training data. Our work demonstrates how foundation vision models can help in solving complex 3D geometric problems such as symmetry detection.