On Robustness of Lane Detection Models to Physical-World Adversarial Attacks in Autonomous Driving

After the 2017 TuSimple Lane Detection Challenge, its evaluation based on accuracy and F1 score has become the de facto standard to measure the performance of lane detection methods. In this work, we conduct the first large-scale empirical study to evaluate the robustness of state-of-the-art lane detection methods under physical-world adversarial attacks in autonomous driving. We evaluate 4 major types of lane detection approaches with the conventional evaluation and end-to-end evaluation in autonomous driving scenarios, and then discuss the security proprieties of each lane detection model. We demonstrate that the conventional evaluation fails to reflect the robustness in end-to-end autonomous driving scenarios. Our results show that the most robust model on the conventional metrics is the least robust in the end-to-end evaluation. Although the competition dataset and its metrics have played a substantial role in developing performant lane detection methods along with the rapid development of deep neural networks, the conventional evaluation is becoming obsolete and the gap between the metrics and practicality is critical. We hope that our study will help the community make further progress in building a more comprehensive framework to evaluate lane detection models.

[1]  Yoshua Bengio,et al.  Algorithms for Hyper-Parameter Optimization , 2011, NIPS.

[2]  Huanyu Wang,et al.  Ultra Fast Structure-aware Deep Lane Detection , 2020, ECCV.

[3]  Chen Change Loy,et al.  Learning Lightweight Lane Detection CNNs by Self Attention Distillation , 2019, 2019 IEEE/CVF International Conference on Computer Vision (ICCV).

[4]  Deng Cai,et al.  RESA: Recurrent Feature-Shift Aggregator for Lane Detection , 2021, AAAI.

[5]  Duen Horng Chau,et al.  ShapeShifter: Robust Physical Adversarial Attack on Faster R-CNN Object Detector , 2018, ECML/PKDD.

[6]  Suman Jana,et al.  DeepTest: Automated Testing of Deep-Neural-Network-Driven Autonomous Cars , 2017, 2018 IEEE/ACM 40th International Conference on Software Engineering (ICSE).

[7]  Satish Chandra,et al.  Identification of Free Flowing Vehicles on Two Lane Intercity Highways under Heterogeneous Traffic condition , 2017 .

[8]  Lujo Bauer,et al.  Accessorize to a Crime: Real and Stealthy Attacks on State-of-the-Art Face Recognition , 2016, CCS.

[9]  Luc Van Gool,et al.  Towards End-to-End Lane Detection: an Instance Segmentation Approach , 2018, 2018 IEEE Intelligent Vehicles Symposium (IV).

[10]  Logan Engstrom,et al.  Black-box Adversarial Attacks with Limited Queries and Information , 2018, ICML.

[11]  Yue Zhao,et al.  Seeing isn't Believing: Practical Adversarial Attack Against Object Detectors , 2018 .

[12]  Junjie Shen,et al.  Dirty Road Can Attack: Security of Deep Learning based Automated Lane Centering under Physical-World Attack , 2020, USENIX Security Symposium.

[13]  Aleksander Madry,et al.  On Adaptive Attacks to Adversarial Example Defenses , 2020, NeurIPS.

[14]  Jonah Philion,et al.  FastDraw: Addressing the Long Tail of Lane Detection by Adapting a Sequential Prediction Network , 2019, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Chunxiao Liu,et al.  Inter-Region Affinity Distillation for Road Marking Segmentation , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[16]  Ming-Wei Chang,et al.  BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , 2019, NAACL.

[17]  Bernhard P. Wrobel,et al.  Multiple View Geometry in Computer Vision , 2001 .

[18]  J. L. Testud,et al.  Paper: Model predictive heuristic control , 1978 .

[19]  Alberto Ferreira de Souza,et al.  PolyLaneNet: Lane Estimation via Deep Polynomial Regression , 2020, 2020 25th International Conference on Pattern Recognition (ICPR).

[20]  Jonathon Shlens,et al.  Explaining and Harnessing Adversarial Examples , 2014, ICLR.

[21]  Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles , 2022 .

[22]  Eder Santana,et al.  A Commute in Data: The comma2k19 Dataset , 2018, ArXiv.

[23]  Atul Prakash,et al.  Query-Efficient Physical Hard-Label Attacks on Deep Learning Visual Classification , 2020, ArXiv.

[24]  Cristina Nita-Rotaru,et al.  Are Self-Driving Cars Secure? Evasion Attacks Against Deep Neural Networks for Steering Angle Prediction , 2019, 2019 IEEE Security and Privacy Workshops (SPW).

[25]  Kai Ma,et al.  Med3D: Transfer Learning for 3D Medical Image Analysis , 2019, ArXiv.

[26]  Seungwoo Yoo,et al.  End-to-End Lane Marker Detection via Row-wise Classification , 2020, 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[27]  Jian Yang,et al.  Line-CNN: End-to-End Traffic Line Detection With Line Proposal Unit , 2020, IEEE Transactions on Intelligent Transportation Systems.

[28]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[29]  Samy Bengio,et al.  Adversarial examples in the physical world , 2016, ICLR.

[30]  Joan Bruna,et al.  Intriguing properties of neural networks , 2013, ICLR.

[31]  Junfeng Yang,et al.  DeepXplore: Automated Whitebox Testing of Deep Learning Systems , 2017, SOSP.

[32]  Francesco Borrelli,et al.  Kinematic and dynamic vehicle models for autonomous driving control design , 2015, 2015 IEEE Intelligent Vehicles Symposium (IV).

[33]  David A. Wagner,et al.  Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples , 2018, ICML.

[34]  Peng Liu,et al.  A review of lane detection methods based on deep learning , 2021, Pattern Recognit..

[35]  Yuval Elovici,et al.  Phantom of the ADAS: Phantom Attacks on Driver-Assistance Systems , 2020, IACR Cryptol. ePrint Arch..

[36]  Ronen Lerner,et al.  Recent progress in road and lane detection: a survey , 2012, Machine Vision and Applications.

[37]  Takenao Ohkawa,et al.  Vehicle Detection Based on Perspective Transformation Using Rear-View Camera , 2011 .

[38]  Rajesh Rajamani,et al.  Vehicle dynamics and control , 2005 .

[39]  Dawn Song,et al.  Physical Adversarial Examples for Object Detectors , 2018, WOOT @ USENIX Security Symposium.

[40]  Logan Engstrom,et al.  Synthesizing Robust Adversarial Examples , 2017, ICML.

[41]  Xiaogang Wang,et al.  Spatial As Deep: Spatial CNN for Traffic Scene Understanding , 2017, AAAI.

[42]  Thiago Oliveira-Santos,et al.  Keep your Eyes on the Lane: Real-time Attention-guided Lane Detection , 2020, 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[43]  Zheng Xu,et al.  Learning to Cluster for Proposal-Free Instance Segmentation , 2018, 2018 International Joint Conference on Neural Networks (IJCNN).

[44]  Lei Xue,et al.  Too Good to Be Safe: Tricking Lane Detection in Autonomous Driving with Crafted Perturbations , 2021, USENIX Security Symposium.