EUVIMEDEUVIMED entdecken
Crossref

CLEAN: Category Knowledge-Driven Compression Framework for Efficient 3D Object Detection

Haonan Zhang, Longjun Liu, Fei Hui, Bo Zhang, Hengmin Zhang, Zhiyuan Zha

IEEE Transactions on Pattern Analysis and Machine Intelligence · 2025-10 · 47 (10) · 8740-8755

DOI: 10.1109/tpami.2025.3582706 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025-10. Verfasst von Haonan Zhang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 66 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Fundamental Research Funds for the Central Universities, National Basic Strengthen Research Program of ReRAM.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Deep neural networks (DNNs) are potent in LiDAR-based 3D object detection (LiDAR-3DOD), yet their deployment remains daunting due to their cumbersome parameters and computations. Knowledge distillation (KD) is promising for compressing DNNs in LiDAR-3DOD. However, most existing KD methods transfer inadequate knowledge between homogeneous detectors, and do not thoroughly explore optimal student architectures, resulting in insufficient gains for compact student detectors. To this end, we propose a category knowledge-driven compression framework to achieve efficient LiDAR-based 3D detectors. Firstly, we distill knowledge from two-stage teacher detectors to one-stage student detectors, overcoming the limitations of homogeneous pairs. To conduct KD in these heterogeneous pairs, we explore the gap between heterogeneous detectors, and introduce category knowledge-driven KD (CaKD), which includes both student-oriented distillation and two-stage-oriented label assignment distillation. Secondly, to search for the optimal architecture of compact student detectors, we introduce a masked category knowledge-driven structured pruning scheme. This scheme evaluates filter importance by analyzing the changes in category predictions related to foreground regions before and after filter removal, and prunes the less important filters accordingly. Finally, we propose a modified IoU-aware redundancy elimination module to remove redundant false positive samples, thereby further improving the accuracy of detectors. Experiments on various point cloud datasets demonstrate that our method delivers impressive results. For example, on KITTI, several compressed one-stage detectors outperform two-stage detectors in both efficiency and accuracy. Besides, on WOD-mini, our framework reduces the memory footprint of CenterPoint by 5.2× and improves the L2 mAPH by 0.55$\%$%.

Förderung

  • National Natural Science Foundation of China
  • Fundamental Research Funds for the Central Universities
  • National Basic Strengthen Research Program of ReRAM
  • Open Project Program of Shanghai Key Laboratory of Data Science

Zitieren

APA

Zhang, H., Liu, L., Hui, F., Zhang, B., Zhang, H., & Zha, Z. (2025). CLEAN: Category Knowledge-Driven Compression Framework for Efficient 3D Object Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(10), 8740-8755. https://doi.org/10.1109/tpami.2025.3582706

Vancouver

Zhang H, Liu L, Hui F, Zhang B, Zhang H, Zha Z. CLEAN: Category Knowledge-Driven Compression Framework for Efficient 3D Object Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2025;47(10):8740-8755. doi:10.1109/tpami.2025.3582706

BibTeX
@article{Zhang2025,
  author = {Zhang, Haonan and Liu, Longjun and Hui, Fei and Zhang, Bo and Zhang, Hengmin and Zha, Zhiyuan},
  title = {CLEAN: Category Knowledge-Driven Compression Framework for Efficient 3D Object Detection},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2025},
  volume = {47},
  number = {10},
  pages = {8740--8755},
  doi = {10.1109/tpami.2025.3582706}
}
Kostenlos starten →