Advanced Active Learning Strategies for Object Detection
🚗 IEEE Intelligent Vehicles Symposium (IV) 2020
Overview
Labeling data for autonomous driving is slow, expensive, and doesn’t scale. What if the model could tell you exactly which samples are worth labeling?
We propose advanced active learning strategies for 2D and 3D object detection using ensemble-based uncertainty estimation. Combined with Continuous Training strategies and careful handling of class imbalance, our approach dramatically cuts labeling costs while maintaining detection performance.
Key Results
⏱️ ~55% time savings in the training pipeline
📉 ~30% data savings for 2D object detection
🎯 35% labeling effort reduction for 3D object detection
⚖️ Analysis of active learning interactions with class imbalance and class weighting
Key Contributions
🎲 Ensemble-based uncertainty — novel uncertainty estimation methods tailored for 2D and 3D detection
🔄 Continuous Training — alleviates growing training times across active learning cycles
🏭 Industry-ready — validated for automotive object detection use cases
Why It Matters
This work demonstrated early on that active learning is not just a theoretical concept — it delivers real, measurable savings for automotive perception, making it practical for production-scale autonomous driving development.
Future self-driving cars must be able to perceive and understand their surroundings. Deep learning based approaches promise to solve the perception problem but require a large amount of manually labeled training data. Active learning is a training procedure in which the model itself selects interesting samples for labeling based on their uncertainty, with substantially less data required for training. Recent research in active learning has mostly focused on the simple image classification task. In this paper, we propose novel methods to estimate sample uncertainties for 2D and 3D object detection using Ensembles. We moreover evaluate different training strategies including Continuous Training to alleviate increasing training times introduced by the active learning cycle. Finally, we investigate the effects of active learning on imbalanced datasets and possible interactions with class weighting. Experiment results show both increased time saving around 55% and data saving rates of around 30%. For the 3D object detection task, we show that our proposed uncertainty estimation method is valid, saving 35% of labeling efforts and thus is ready for application for automotive object detection use cases.
@inproceedings{Schmidt2020,title={Advanced Active Learning Strategies for Object Detection},author={Schmidt, Sebastian and Rao, Qing and Tatsch, Julian and Knoll, Alois},year={2020},booktitle={Proceedings of the IEEE Intelligent Vehicles Symposium (IV)},issue={Iv},_pages={871-876},url={https://mediatum.ub.tum.de/doc/1585225},}