ShanghaiTech University Knowledge Management System
Point-to-Voxel Knowledge Distillation for LiDAR Semantic Segmentation | |
2022 | |
会议录名称 | PROCEEDINGS OF THE IEEE COMPUTER SOCIETY CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION |
ISSN | 1063-6919 |
卷号 | 2022-June |
页码 | 8469-8478 |
发表状态 | 已发表 |
DOI | 10.1109/CVPR52688.2022.00829 |
摘要 | This article addresses the problem of distilling knowledge from a large teacher model to a slim student network for LiDAR semantic segmentation. Directly employing previous distillation approaches yields inferior results due to the intrinsic challenges of point cloud, i.e., sparsity, randomness and varying density. To tackle the aforementioned problems, we propose the Point-to-Voxel Knowledge Distillation (PVD), which transfers the hidden knowledge from both point level and voxel level. Specifically, we first leverage both the pointwise and voxelwise output distillation to complement the sparse supervision signals. Then, to better exploit the structural information, we divide the whole point cloud into several supervoxels and design a difficulty-aware sampling strategy to more frequently sample supervoxels containing less-frequent classes and faraway objects. On these supervoxels, we propose inter-point and inter-voxel affinity distillation, where the similarity information between points and voxels can help the student model better capture the structural information of the surrounding environment. We conduct extensive experiments on two popular LiDAR segmentation benchmarks, i.e., nuScenes [3] and SemanticKITTI [1]. On both benchmarks, our PVD-consistently outperforms previous distillation approaches by a large margin on three representative backbones, i.e., Cylinder3D [36], [37], SPVNAS [25] and MinkowskiNet [5]. Notably, on the challenging nuScenes and SemanticKITTI datasets, our method can achieve roughly 75% MACs reduction and 2× speedup on the competitive Cylinder3D model and rank 1st on the SemanticKITTI leaderboard among all published algorithms11https://competitions.codalab.org/competitions/20331#results (single-scan competition) till 2021-11-18 04:00 Pacific Time, and our method is termed Point-Voxel-KD. Our method (PV-KD) ranks 3rd on the multi-scan challenge till 2021-12-1 00:00 Pacific Time. Our code is available at https://github.com/cardwing/Codes-for-PVKD. © 2022 IEEE. |
会议名称 | 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022 |
出版地 | 10662 LOS VAQUEROS CIRCLE, PO BOX 3014, LOS ALAMITOS, CA 90720-1264 USA |
会议地点 | New Orleans, LA, United states |
会议日期 | June 19, 2022 - June 24, 2022 |
URL | 查看原文 |
收录类别 | EI ; CPCI-S |
语种 | 英语 |
WOS研究方向 | Computer Science ; Imaging Science & Photographic Technology |
WOS类目 | Computer Science, Artificial Intelligence ; Imaging Science & Photographic Technology |
WOS记录号 | WOS:000870759101051 |
出版者 | IEEE Computer Society |
EI入藏号 | 20224613120193 |
原始文献类型 | Conference article (CA) |
来源库 | IEEE |
引用统计 | 正在获取...
|
文献类型 | 会议论文 |
条目标识符 | https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/248931 |
专题 | 信息科学与技术学院_PI研究组_马月昕 |
通讯作者 | Hou, Yuenan |
作者单位 | 1.Shanghai AI Lab, Shanghai, Peoples R China 2.Chinese Univ Hong Kong, Hong Kong, Peoples R China 3.ShanghaiTech Univ, Shanghai, Peoples R China 4.Nanyang Technol Univ, S Lab, Singapore, Singapore |
推荐引用方式 GB/T 7714 | Hou, Yuenan,Zhu, Xinge,Ma, Yuexin,et al. Point-to-Voxel Knowledge Distillation for LiDAR Semantic Segmentation[C]. 10662 LOS VAQUEROS CIRCLE, PO BOX 3014, LOS ALAMITOS, CA 90720-1264 USA:IEEE Computer Society,2022:8469-8478. |
条目包含的文件 | 下载所有文件 | |||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 |
修改评论
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。