Adversarial Attacks on Multi-task Visual Perception for Autonomous Driving
Authors:
Preprint
Abstract
Deep neural networks (DNNs) have accomplished impressive success in various applications, including autonomous driving perception tasks, in recent years. On the other hand, current deep neural networks are easily fooled by adversarial attacks. This vulnerability raises significant concerns, particularly in safety-critical applications. As a result, research into attacking and defending DNNs has gained much coverage. In this work, detailed adversarial attacks are applied on a diverse multi-task visual perception deep network across distance estimation, semantic segmentation, motion detection, and object detection. The experiments consider both white and black box attacks for targeted and un-targeted cases, while attacking a task and inspecting the effect on all the others, in addition to inspecting the effect of applying a simple defense method. We conclude this paper by comparing and discussing the experimental results, proposing insights and future work. The visualizations of the attacks are available at https://youtu.be/6AixN90budY.Keywords
All Related Versions
This publication has 6 references indexed in Scilit:
- Inspect, Understand, Overcome: A Survey of Practical Methods for AI Safety2022
- RoadEdgeNet: Road Edge Detection System Using Surround View Camera ImagesElectronic Imaging, 2021
- Let’s Get Dirty: GAN Based Data Augmentation for Camera Lens Soiling Detection in Autonomous Driving2021
- Multitask Learning Strengthens Adversarial Robustness2020
- Dynamic Task Weighting Methods for Multi-task Networks in Autonomous Driving Systems2020
- TiledSoilingNet: Tile-level Soiling Detection on Automotive Surround-view Cameras Using Coverage Metric2020