Shape-Biased Domain Generalization via Shock Graph Embeddings

TL;DR

This paper proposes using shock graph embeddings to explicitly represent shape content in images, improving domain generalization of neural networks by reducing texture bias and enhancing shape awareness.

Contribution

It introduces a novel shape representation via shock graphs combined with GNNs, outperforming traditional CNNs in domain generalization without relying on appearance features.

Findings

Shape-based shock graph embeddings outperform CNNs in domain generalization.

The approach achieves higher accuracy on Colored MNIST, PACS, and VLCS datasets.

Explicit shape representation reduces texture bias and improves robustness.

Abstract

There is an emerging sense that the vulnerability of Image Convolutional Neural Networks (CNN), i.e., sensitivity to image corruptions, perturbations, and adversarial attacks, is connected with Texture Bias. This relative lack of Shape Bias is also responsible for poor performance in Domain Generalization (DG). The inclusion of a role of shape alleviates these vulnerabilities and some approaches have achieved this by training on negative images, images endowed with edge maps, or images with conflicting shape and texture information. This paper advocates an explicit and complete representation of shape using a classical computer vision approach, namely, representing the shape content of an image with the shock graph of its contour map. The resulting graph and its descriptor is a complete representation of contour content and is classified using recent Graph Neural Network (GNN) methods. The experimental results on three domain shift datasets, Colored MNIST, PACS, and VLCS demonstrate that even without using appearance the shape-based approach exceeds classical Image CNN based methods in domain generalization.