Unleashing the Potential of Unlabeled Data: Bidirectional Collaborative Semi-Supervised Active Learning for 3D Object Detection

Zengran Wang, Yanan Zhang, Jiaxin Chen, Di Huang

Keywords: Autonomous Driving, 3D Object Detection, Active Learning, Semi-Supervised Learning

Abstract:

To address the annotation burden in LiDAR-based 3D object detection, active learning (AL) methods offer a promising solution. However, traditional active learning approaches solely rely on labeled data to train an initial model for data selection, overlooking the potential of leveraging unlabeled data. Recently, attempts to integrate semi-supervised learning (SSL) into AL with the goal of leveraging unlabeled data have faced challenges in effectively resolving the conflict between the two paradigms, resulting in less satisfactory performance.
To tackle this conflict, we propose a Bidirectional Collaborative Semi-Supervised Active Learning framework, dubbed as BC-SSAL. Specifically, from the perspective of SSL, we propose a Collaborative PseudoScene Pre-training (CPSP) method that effectively learns from unlabeled data without introducing adverse effects. From the perspective of AL, we design a Collaborative Active Learning (CAL) method tailored for outdoor LiDAR scenes, which complements the uncertainty and diversity methods by model cascading, alleviating the dilemma of sampling rare classes. Extensive experiments conducted on KITTI and Waymo demonstrate the effectiveness of our BC-SSAL. Especially, on the KITTI dataset, utilizing only 2\% labeled data, BC-SSAL can achieve comparable performance to the model trained on the full set.

Primary Area: applications to computer vision, audio, language, and other modalities

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 4542