Dual Geometric Graph Network (DG2N) -- Iterative network for deformable shape alignment
This work provides an incremental improvement in deformable shape alignment, which is a challenging computer vision task, especially for inter-class scenarios with non-isometric deformations.
This paper addresses the problem of aligning non-rigid geometric models, particularly for inter-class alignment where non-isometric deformations are common. The authors propose an iterative network based on a dual graph structure, achieving state-of-the-art results on stretchable domain alignment for meshes and point clouds.
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.