Research on PointPillars Algorithm Based on Feature-Enhanced Backbone Network
In the industrial field, the 3D target detection algorithm PointPillars has gained popularity. Improving target detection accuracy while maintaining high efficiency has been a significant challenge. To address the issue of low target detection accuracy in the PointPillars 3D target detection algorit...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-03-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/13/7/1233 |
| _version_ | 1797212704748863488 |
|---|---|
| author | Xiaoning Shu Liang Zhang |
| author_facet | Xiaoning Shu Liang Zhang |
| author_sort | Xiaoning Shu |
| collection | DOAJ |
| description | In the industrial field, the 3D target detection algorithm PointPillars has gained popularity. Improving target detection accuracy while maintaining high efficiency has been a significant challenge. To address the issue of low target detection accuracy in the PointPillars 3D target detection algorithm, this paper proposes an algorithm based on feature enhancement to improve the backbone network. The algorithm enhances preliminary feature information of the backbone network by modifying it based on PointPillars with the aid of channel attention and spatial attention mechanisms. To address the inefficiency caused by the excessive number of subsampled parameters in PointPillars, FasterNet (a lightweight and efficient feature extraction network) is utilized for down-sampling and forming different scale feature maps. To prevent the loss and blurring of extracted features resulting from the use of inverse convolution, we utilize the lightweight and efficient up-sampling modules Carafe and Dysample for adjusting resolution. Experimental results indicate improved accuracy under all difficulties of the KITTI dataset, demonstrating the superiority of the algorithm over PointPillars. |
| first_indexed | 2024-04-24T10:46:37Z |
| format | Article |
| id | doaj.art-658837b1a950486f8d5bddc89aa6aee0 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-04-24T10:46:37Z |
| publishDate | 2024-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-658837b1a950486f8d5bddc89aa6aee02024-04-12T13:17:08ZengMDPI AGElectronics2079-92922024-03-01137123310.3390/electronics13071233Research on PointPillars Algorithm Based on Feature-Enhanced Backbone NetworkXiaoning Shu0Liang Zhang1Qingdao Computing Technology Research Institute, Xidian University, Qingdao 266071, ChinaSchool of Computer Science and Technology, Xidian University, Xi’an 710119, ChinaIn the industrial field, the 3D target detection algorithm PointPillars has gained popularity. Improving target detection accuracy while maintaining high efficiency has been a significant challenge. To address the issue of low target detection accuracy in the PointPillars 3D target detection algorithm, this paper proposes an algorithm based on feature enhancement to improve the backbone network. The algorithm enhances preliminary feature information of the backbone network by modifying it based on PointPillars with the aid of channel attention and spatial attention mechanisms. To address the inefficiency caused by the excessive number of subsampled parameters in PointPillars, FasterNet (a lightweight and efficient feature extraction network) is utilized for down-sampling and forming different scale feature maps. To prevent the loss and blurring of extracted features resulting from the use of inverse convolution, we utilize the lightweight and efficient up-sampling modules Carafe and Dysample for adjusting resolution. Experimental results indicate improved accuracy under all difficulties of the KITTI dataset, demonstrating the superiority of the algorithm over PointPillars.https://www.mdpi.com/2079-9292/13/7/1233target detectionPointPillarsFasterNetup-samplingattention mechanism |
| spellingShingle | Xiaoning Shu Liang Zhang Research on PointPillars Algorithm Based on Feature-Enhanced Backbone Network Electronics target detection PointPillars FasterNet up-sampling attention mechanism |
| title | Research on PointPillars Algorithm Based on Feature-Enhanced Backbone Network |
| title_full | Research on PointPillars Algorithm Based on Feature-Enhanced Backbone Network |
| title_fullStr | Research on PointPillars Algorithm Based on Feature-Enhanced Backbone Network |
| title_full_unstemmed | Research on PointPillars Algorithm Based on Feature-Enhanced Backbone Network |
| title_short | Research on PointPillars Algorithm Based on Feature-Enhanced Backbone Network |
| title_sort | research on pointpillars algorithm based on feature enhanced backbone network |
| topic | target detection PointPillars FasterNet up-sampling attention mechanism |
| url | https://www.mdpi.com/2079-9292/13/7/1233 |
| work_keys_str_mv | AT xiaoningshu researchonpointpillarsalgorithmbasedonfeatureenhancedbackbonenetwork AT liangzhang researchonpointpillarsalgorithmbasedonfeatureenhancedbackbonenetwork |