Shape-Texture Debiased Neural Network Training

TL;DR

This paper introduces a simple data augmentation method that reduces shape-texture bias in neural networks, leading to improved accuracy and robustness across multiple image recognition benchmarks.

Contribution

The authors propose a novel shape-texture debiased training algorithm that combines conflicting cues and dual supervision, enhancing model generalization and robustness.

Findings

Improved accuracy on ImageNet (+1.2%) and adversarial robustness (+14.4%).

Enhanced performance on ImageNet-A, ImageNet-C, and Stylized-ImageNet datasets.

Compatible with other data augmentation techniques like Mixup and CutMix.

Abstract

Shape and texture are two prominent and complementary cues for recognizing objects. Nonetheless, Convolutional Neural Networks are often biased towards either texture or shape, depending on the training dataset. Our ablation shows that such bias degenerates model performance. Motivated by this observation, we develop a simple algorithm for shape-texture debiased learning. To prevent models from exclusively attending on a single cue in representation learning, we augment training data with images with conflicting shape and texture information (eg, an image of chimpanzee shape but with lemon texture) and, most importantly, provide the corresponding supervisions from shape and texture simultaneously. Experiments show that our method successfully improves model performance on several image recognition benchmarks and adversarial robustness. For example, by training on ImageNet, it helps ResNet-152 achieve substantial improvements on ImageNet (+1.2%), ImageNet-A (+5.2%), ImageNet-C (+8.3%) and Stylized-ImageNet (+11.1%), and on defending against FGSM adversarial attacker on ImageNet (+14.4%). Our method also claims to be compatible with other advanced data augmentation strategies, eg, Mixup, and CutMix. The code is available here: https://github.com/LiYingwei/ShapeTextureDebiasedTraining.