Affine Disentangled GAN for Interpretable and Robust AV Perception

TL;DR

This paper introduces ADIS-GAN, a deep learning model designed to improve autonomous vehicle perception by being robust against affine transformations and adversarial attacks, while also providing interpretable features.

Contribution

The paper presents a novel GAN architecture that handles both affine transformations and adversarial attacks simultaneously, enhancing robustness and interpretability in AV perception systems.

Findings

Achieves over 98% accuracy within 30° rotation on MNIST

Over 90% accuracy against FGSM and PGD attacks

Generates interpretable information like rotation angle and scaling factor

Abstract

Autonomous vehicles (AV) have progressed rapidly with the advancements in computer vision algorithms. The deep convolutional neural network as the main contributor to this advancement has boosted the classification accuracy dramatically. However, the discovery of adversarial examples reveals the generalization gap between dataset and the real world. Furthermore, affine transformations may also confuse computer vision based object detectors. The degradation of the perception system is undesirable for safety critical systems such as autonomous vehicles. In this paper, a deep learning system is proposed: Affine Disentangled GAN (ADIS-GAN), which is robust against affine transformations and adversarial attacks. It is demonstrated that conventional data augmentation for affine transformation and adversarial attacks are orthogonal, while ADIS-GAN can handle both attacks at the same time. Useful information such as image rotation angle and scaling factor are also generated in ADIS-GAN. On MNIST dataset, ADIS-GAN can achieve over 98 percent classification accuracy within 30 degrees rotation, and over 90 percent classification accuracy against FGSM and PGD adversarial attack.