Distribution Discrepancy and Feature Heterogeneity for Active 3D Object Detection
Not on the map. Why?
Abstract (arXiv)
LiDAR-based 3D object detection is a critical technology for the development of autonomous driving and robotics. However, the high cost of data annotation limits its advancement. We propose a novel and effective active learning (AL) method called Distribution Discrepancy and Feature Heterogeneity (DDFH), which simultaneously considers geometric features and model embeddings, assessing information from both the instance-level and frame-level perspectives. Distribution Discrepancy evaluates the difference and novelty of instances within the unlabeled and labeled distributions, enabling the model to learn efficiently with limited data. Feature Heterogeneity ensures the heterogeneity of intra-frame instance features, maintaining feature diversity while avoiding redundant or similar instances, thus minimizing annotation costs. Finally, multiple indicators are efficiently aggregated using Quantile Transform, providing a unified measure of informativeness. Extensive experiments demonstrate that DDFH outperforms the current state-of-the-art (SOTA) methods on the KITTI and Waymo datasets, effectively reducing the bounding box annotation cost by 56.3% and showing robustness when working with both one-stage and two-stage models.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 29 of its 44 references that are in the map. First posted on arXiv 9 Sep 2024.
Builds on
29 of 44 references.
References · 29
- 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
- 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
- 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,436
- 2019PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detectioncited by 890
- 2017Deep Bayesian Active Learning with Image Datacited by 812
- 2019Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Boundscited by 602
- 2019Learning Loss for Active Learningcited by 335
- 2021Voxel Transformer for 3D Object Detectioncited by 247
- 2021PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detectioncited by 232
- 2020Contextual Diversity for Active Learningcited by 139
- 2019ViewAL: Active Learning with Viewpoint Entropy for Semantic Segmentationcited by 95
- 2021Gone Fishing: Neural Active Learning with Fisher Embeddingscited by 88
17 more
- 2021Active Learning for Deep Object Detection via Probabilistic Modelingcited by 79
- 2021Multiple instance active learning for object detectioncited by 77
- 2019BayesOD: A Bayesian Approach for Uncertainty Estimation in Deep Object Detectorscited by 68
- 2022Active Learning by Feature Mixingcited by 63
- 2022Entropy-based Active Learning for Object Detection with Progressive Diversity Constraintcited by 53
- 2018Localization-Aware Active Learning for Object Detectioncited by 48
- 2019Deep Active Learning for Efficient Training of a LiDAR 3D Object Detectorcited by 45
- 2021Diminishing Uncertainty within the Training Pool: Active Learning for Medical Image Segmentationcited by 44
- 2022Plug and Play Active Learning for Object Detectioncited by 28
- 2023Exploring Active 3D Object Detection from a Generalization Perspectivecited by 26
- 2021MCDAL: Maximum Classifier Discrepancy for Active Learningcited by 18
- 2024ActiveAnno3D -- An Active Learning Framework for Multi-Modal 3D Object Detectioncited by 17
- 2020DIRL: Domain-Invariant Representation Learning for Sim-to-Real Transfercited by 13
- 2023KECOR: Kernel Coding Rate Maximization for Active 3D Object Detectioncited by 13
- 20234D-Former: Multimodal 4D Panoptic Segmentationcited by 12
- 2023DORT: Modeling Dynamic Objects in Recurrent for Multi-Camera 3D Object Detection and Trackingcited by 11
- 2023LabelFormer: Object Trajectory Refinement for Offboard Perception from LiDAR Point Cloudscited by 7
Built on by
Citing papers per year: 2025: 1.
1 paper.