Real-Time Fully Unsupervised Domain Adaptation for Lane Detection in Autonomous Driving
Abstract: While deep neural networks are being utilized heavily for autonomous driving, they need to be adapted to new unseen environmental conditions for which they were not trained. We focus on a safety critical application of lane detection, and propose a lightweight, fully unsupervised, real-time adaptation approach that only adapts the batch-normalization parameters of the model. We demonstrate that our technique can perform inference, followed by on-device adaptation, under a tight constraint of 30 FPS on Nvidia Jetson Orin. It shows similar accuracy (avg. of 92.19%) as a state-of-the-art semi-supervised adaptation algorithm but which does not support real-time adaptation.
- Carla: An open urban driving simulator. In Conference on robot learning (CoRL), pages 1–16. PMLR, 2017.
- Carlane: A lane detection benchmark for unsupervised domain adaptation from simulation to multiple real-world domains. Advances in Neural Information Processing Systems (NeurIPS), 35:4046–4058, 2022.
- Ultra fast structure-aware deep lane detection. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXIV 16, pages 276–291. Springer, 2020.
- TuSimple. TuSimple-benchmark. https://github.com/TuSimple/tusimple-benchmark/tree/master/doc/lane_detection.
- Tent: Fully test-time adaptation by entropy minimization. International Conference on Learning Representations (ICLR), 2021.
- D. McCarthy. Vision Navigates Obstacles on the Road to Autonomous Vehicles. https://www.automate.org/industry-insights/vision-navigates-obstacles-on-the-road-to-autonomous-vehicles.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.