Dual Geometric Graph Network (DG2N) -- Iterative network for deformable shape alignment

TL;DR

This paper introduces DG2N, an iterative dual graph network for deformable shape alignment that effectively handles non-isometric deformations, outperforming existing methods on stretchable domain alignments.

Contribution

The paper presents a novel dual graph network architecture with unrolling for non-rigid shape alignment, addressing limitations of prior DNN-based methods in inter-class deformations.

Findings

Achieves state-of-the-art results on stretchable domain alignment

Provides a rapid and stable solution for mesh and point cloud alignment

Effectively handles non-isometric deformations in shape matching

Abstract

We provide a novel new approach for aligning geometric models using a dual graph structure where local features are mapping probabilities. Alignment of non-rigid structures is one of the most challenging computer vision tasks due to the high number of unknowns needed to model the correspondence. We have seen a leap forward using DNN models in template alignment and functional maps, but those methods fail for inter-class alignment where nonisometric deformations exist. Here we propose to rethink this task and use unrolling concepts on a dual graph structure - one for a forward map and one for a backward map, where the features are pulled back matching probabilities from the target into the source. We report state of the art results on stretchable domains alignment in a rapid and stable solution for meshes and cloud of points.