Geometric Point Attention Transformer for 3D Shape Reassembly

TL;DR

This paper introduces GPAT, a transformer-based network that improves 3D shape reassembly by better capturing geometric relationships between parts through an innovative attention mechanism and iterative refinement.

Contribution

The paper proposes the Geometric Point Attention Transformer (GPAT), integrating global and local geometric features with a recycling scheme for iterative pose refinement in 3D shape assembly.

Findings

Outperforms previous methods in pose estimation accuracy

Achieves high alignment accuracy in shape reassembly

Effective in both semantic and geometric assembly tasks

Abstract

Shape assembly, which aims to reassemble separate parts into a complete object, has gained significant interest in recent years. Existing methods primarily rely on networks to predict the poses of individual parts, but often fail to effectively capture the geometric interactions between the parts and their poses. In this paper, we present the Geometric Point Attention Transformer (GPAT), a network specifically designed to address the challenges of reasoning about geometric relationships. In the geometric point attention module, we integrate both global shape information and local pairwise geometric features, along with poses represented as rotation and translation vectors for each part. To enable iterative updates and dynamic reasoning, we introduce a geometric recycling scheme, where each prediction is fed into the next iteration for refinement. We evaluate our model on both the semantic and geometric assembly tasks, showing that it outperforms previous methods in absolute pose estimation, achieving accurate pose predictions and high alignment accuracy.