Topology-Guided Knowledge Distillation for Efficient Point Cloud Processing
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912373066956800 |
|---|---|
| author | Hai, Luu Tung Le, Thinh D. Ding, Zhicheng Tian, Qing Hy, Truong-Son |
| author_facet | Hai, Luu Tung Le, Thinh D. Ding, Zhicheng Tian, Qing Hy, Truong-Son |
| contents | Point cloud processing has gained significant attention due to its critical role in applications such as autonomous driving and 3D object recognition. However, deploying high-performance models like Point Transformer V3 in resource-constrained environments remains challenging due to their high computational and memory demands. This work introduces a novel distillation framework that leverages topology-aware representations and gradient-guided knowledge distillation to effectively transfer knowledge from a high-capacity teacher to a lightweight student model. Our approach captures the underlying geometric structures of point clouds while selectively guiding the student model's learning process through gradient-based feature alignment. Experimental results in the Nuscenes, SemanticKITTI, and Waymo datasets demonstrate that the proposed method achieves competitive performance, with an approximately 16x reduction in model size and a nearly 1.9x decrease in inference time compared to its teacher model. Notably, on NuScenes, our method achieves state-of-the-art performance among knowledge distillation techniques trained solely on LiDAR data, surpassing prior knowledge distillation baselines in segmentation performance. Our implementation is available publicly at:
https://github.com/HySonLab/PointDistill |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_08101 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Topology-Guided Knowledge Distillation for Efficient Point Cloud Processing Hai, Luu Tung Le, Thinh D. Ding, Zhicheng Tian, Qing Hy, Truong-Son Computer Vision and Pattern Recognition Machine Learning Point cloud processing has gained significant attention due to its critical role in applications such as autonomous driving and 3D object recognition. However, deploying high-performance models like Point Transformer V3 in resource-constrained environments remains challenging due to their high computational and memory demands. This work introduces a novel distillation framework that leverages topology-aware representations and gradient-guided knowledge distillation to effectively transfer knowledge from a high-capacity teacher to a lightweight student model. Our approach captures the underlying geometric structures of point clouds while selectively guiding the student model's learning process through gradient-based feature alignment. Experimental results in the Nuscenes, SemanticKITTI, and Waymo datasets demonstrate that the proposed method achieves competitive performance, with an approximately 16x reduction in model size and a nearly 1.9x decrease in inference time compared to its teacher model. Notably, on NuScenes, our method achieves state-of-the-art performance among knowledge distillation techniques trained solely on LiDAR data, surpassing prior knowledge distillation baselines in segmentation performance. Our implementation is available publicly at: https://github.com/HySonLab/PointDistill |
| title | Topology-Guided Knowledge Distillation for Efficient Point Cloud Processing |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2505.08101 |