GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and Segmentation
The processing of point cloud data has become a significant area of research in the modern field of perception. Classification and segmentation are critical tasks in autonomous driving, environmental perception, and digital twins. Algorithms that directly extract features from raw point cloud data h...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2025-01-01
|
| Series: | IET Computers & Digital Techniques |
| Online Access: | http://dx.doi.org/10.1049/cdt2/7934018 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850263498084319232 |
|---|---|
| author | Zourong Long Gen Tan You Wu Hong Yang Chao Ding |
| author_facet | Zourong Long Gen Tan You Wu Hong Yang Chao Ding |
| author_sort | Zourong Long |
| collection | DOAJ |
| description | The processing of point cloud data has become a significant area of research in the modern field of perception. Classification and segmentation are critical tasks in autonomous driving, environmental perception, and digital twins. Algorithms that directly extract features from raw point cloud data have simple architectures, but they are constrained by computational demands and limited efficiency. This makes effective deployment on resource-limited devices challenging. This article introduces GRSNet, an ultra-lightweight algorithm. The principal innovation is a new sampling method named golden ratio sampling (GRS), which generates sampling point indices directly using the golden ratio to subsequently locate the corresponding sampling points. This method efficiently extracts representative points from point cloud data and integrates them into deep networks. Leveraging GRS, this study combines the concepts from GhostNet and self-attention mechanisms to develop a feature extraction module dubbed the SA_Ghost Block, forming the core of GRSNet. Comparative experiments with leading algorithms on established point cloud open-source datasets demonstrate that GRSNet achieves superior performance, maintaining only 0.7 M parameters. |
| format | Article |
| id | doaj-art-1a357cd98e9c406487a3f553524abb7e |
| institution | OA Journals |
| issn | 1751-861X |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Computers & Digital Techniques |
| spelling | doaj-art-1a357cd98e9c406487a3f553524abb7e2025-08-20T01:54:57ZengWileyIET Computers & Digital Techniques1751-861X2025-01-01202510.1049/cdt2/7934018GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and SegmentationZourong Long0Gen Tan1You Wu2Hong Yang3Chao Ding4Chongqing University of TechnologyChongqing University of TechnologyChongqing University of TechnologyChongqing University of TechnologyChongqing University of TechnologyThe processing of point cloud data has become a significant area of research in the modern field of perception. Classification and segmentation are critical tasks in autonomous driving, environmental perception, and digital twins. Algorithms that directly extract features from raw point cloud data have simple architectures, but they are constrained by computational demands and limited efficiency. This makes effective deployment on resource-limited devices challenging. This article introduces GRSNet, an ultra-lightweight algorithm. The principal innovation is a new sampling method named golden ratio sampling (GRS), which generates sampling point indices directly using the golden ratio to subsequently locate the corresponding sampling points. This method efficiently extracts representative points from point cloud data and integrates them into deep networks. Leveraging GRS, this study combines the concepts from GhostNet and self-attention mechanisms to develop a feature extraction module dubbed the SA_Ghost Block, forming the core of GRSNet. Comparative experiments with leading algorithms on established point cloud open-source datasets demonstrate that GRSNet achieves superior performance, maintaining only 0.7 M parameters.http://dx.doi.org/10.1049/cdt2/7934018 |
| spellingShingle | Zourong Long Gen Tan You Wu Hong Yang Chao Ding GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and Segmentation IET Computers & Digital Techniques |
| title | GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and Segmentation |
| title_full | GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and Segmentation |
| title_fullStr | GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and Segmentation |
| title_full_unstemmed | GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and Segmentation |
| title_short | GRSNet: An Ultra-Lightweight Neural Network for 3D Point Cloud Classification and Segmentation |
| title_sort | grsnet an ultra lightweight neural network for 3d point cloud classification and segmentation |
| url | http://dx.doi.org/10.1049/cdt2/7934018 |
| work_keys_str_mv | AT zouronglong grsnetanultralightweightneuralnetworkfor3dpointcloudclassificationandsegmentation AT gentan grsnetanultralightweightneuralnetworkfor3dpointcloudclassificationandsegmentation AT youwu grsnetanultralightweightneuralnetworkfor3dpointcloudclassificationandsegmentation AT hongyang grsnetanultralightweightneuralnetworkfor3dpointcloudclassificationandsegmentation AT chaoding grsnetanultralightweightneuralnetworkfor3dpointcloudclassificationandsegmentation |