자가학습과 지식증류 방법을 활용한LiDAR 3차원 물체 탐지에서의 준지도 도메인 적응
Semi-Supervised Domain Adaptation on LiDAR 3D Object Detection with Self-Training and Knowledge Distillation
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
With the release of numerous open driving datasets, the demand for domain adaptation in perception tasks has increased, particularly when transferring knowledge from rich datasets to novel domains. However, it is difficult to solve the change 1) in the sensor domain caused by heterogeneous LiDAR sensors and 2) in the environmental domain caused by different environmental factors. We overcome domain differences in the semi-supervised setting with 3-stage model parameter training. First, we pre-train the model with the source dataset with object scaling based on statistics of the object size. Then we fine-tine the partially frozen model weights with copy-and-paste augmentation. The 3D points in the box labels are copied from one scene and pasted to the other scenes. Finally, we use the knowledge distillation method to update the student network with a moving average from the teacher network along with a self-training method with pseudo labels. Test-Time Augmentation with varying z values is employed to predict the final results. Our method achieved 3rd place in ECCV 2022 workshop on the 3D Perception for Autonomous Driving challenge.
Abstract
With the release of numerous open driving datasets, the demand for domain adaptation in perception tasks has increased, particularly when transferring knowledge from rich datasets to novel domains. However, it is difficult to solve the change 1) in the sensor domain caused by heterogeneous LiDAR sensors and 2) in the environmental domain caused by different environmental factors. We overcome domain differences in the semi-supervised setting with 3-stage model parameter training. First, we pre-train the model with the source dataset with object scaling based on statistics of the object size. Then we fine-tine the partially frozen model weights with copy-and-paste augmentation. The 3D points in the box labels are copied from one scene and pasted to the other scenes. Finally, we use the knowledge distillation method to update the student network with a moving average from the teacher network along with a self-training method with pseudo labels. Test-Time Augmentation with varying z values is employed to predict the final results. Our method achieved 3rd place in ECCV 2022 workshop on the 3D Perception for Autonomous Driving challenge.
참고문헌 (15)
- Pei Sun, "Scalability in Perception for Autonomous Driving: Waymo Open Dataset", 2020 IEEE, 2020, ~[DOI]
- A Geiger, "Vision meets robotics: The KITTI dataset", The International Journal of Robotics Research, 32(11), pp.1231~1237 (2013), SAGE Publications[DOI]
- Holger Caesar, "nuScenes: A Multimodal Dataset for Autonomous Driving", 2020 IEEE, 2020, ~[DOI]
- J. Houston, "One Thousand and One Hours: Self-driving Motion Prediction Dataset", Conference on Robot Learning (CoRL), pp.409~418 (2020)
- Ming-Fang Chang, "Argoverse: 3D Tracking and Forecasting With Rich Maps", 2019 IEEE, 2019, ~[DOI]
- Pengchuan Xiao, "PandaSet: Advanced Sensor Suite Dataset for Autonomous Driving", 2021 IEEE International Intelligent Transportation Systems Conference (ITSC) (2021), IEEE
- Ricardo Roriz, "Automotive LiDAR Technology: A Survey", IEEE Transactions on Intelligent Transportation Systems, 23(7), pp.6282~6297 (2022), Institute of Electrical and Electronics Engineers (IEEE)[DOI]
- Lu Jie, "A SLAM System with Direct Velocity Estimation for Mechanical and Solid-State LiDARs", Remote Sensing, 14(7), pp.1741~ (2022), MDPI AG[DOI]
- Yan Wang, "Train in Germany, Test in the USA: Making 3D Object Detectors Generalize", 2020 IEEE, 2020, ~[DOI]
- Tianwei Yin, "Center-based 3D Object Detection and Tracking", 2021 IEEE, 2021, ~[DOI]
- Shaoshuai Shi, "PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection", International Journal of Computer Vision, 131(2), pp.531~551 (2022), Springer Science and Business Media LLC[DOI]
- Shaoshuai Shi, "PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection", 2020 IEEE, 2020, ~[DOI]
- S. Ioffe, "Batch Normalization : Accelerating Deep Network Training by Reducing Internal Covariate Shift", International Conference on Machine Learning, Lille, pp.448~456 (2015)
- Yan Yan, "SECOND: Sparsely Embedded Convolutional Detection", Sensors, 18(10), pp.3337~ (2018), MDPI AG[DOI]
- J.-B. Grill, "Advances in Neural Information Processing Systems", pp.21271~21284 (2020), Red Hook