Point cloud semantic segmentation based on local feature fusion and multilayer attention network
Abstract Semantic segmentation from a three‐dimensional point cloud is vital in autonomous driving, computer vision, and augmented reality. However, current semantic segmentation does not effectively use the point cloud's local geometric features and contextual information, essential for improv...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2024-04-01
|
| Series: | IET Computer Vision |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/cvi2.12255 |
| _version_ | 1827284865284833280 |
|---|---|
| author | Junjie Wen Jie Ma Yuehua Zhao Tong Nie Mengxuan Sun Ziming Fan |
| author_facet | Junjie Wen Jie Ma Yuehua Zhao Tong Nie Mengxuan Sun Ziming Fan |
| author_sort | Junjie Wen |
| collection | DOAJ |
| description | Abstract Semantic segmentation from a three‐dimensional point cloud is vital in autonomous driving, computer vision, and augmented reality. However, current semantic segmentation does not effectively use the point cloud's local geometric features and contextual information, essential for improving segmentation accuracy. A semantic segmentation network that uses local feature fusion and a multilayer attention mechanism is proposed to address these challenges. Specifically, the authors designed a local feature fusion module to encode the geometric and feature information separately, which fully leverages the point cloud's feature perception and geometric structure representation. Furthermore, the authors designed a multilayer attention pooling module consisting of local attention pooling and cascade attention pooling to extract contextual information. Local attention pooling is used to learn local neighbourhood information, and cascade attention pooling captures contextual information from deeper local neighbourhoods. Finally, an enhanced feature representation of important information is obtained by aggregating the features from the two deep attention pooling methods. Extensive experiments on large‐scale point‐cloud datasets Stanford 3D large‐scale indoor spaces and SemanticKITTI indicate that authors network shows excellent advantages over existing representative methods regarding local geometric feature description and global contextual relationships. |
| first_indexed | 2024-04-24T10:05:36Z |
| format | Article |
| id | doaj.art-beaa3c9ae24f4ebfbf3c162cd9cb9eb8 |
| institution | Directory Open Access Journal |
| issn | 1751-9632 1751-9640 |
| language | English |
| last_indexed | 2024-04-24T10:05:36Z |
| publishDate | 2024-04-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Computer Vision |
| spelling | doaj.art-beaa3c9ae24f4ebfbf3c162cd9cb9eb82024-04-13T04:15:00ZengWileyIET Computer Vision1751-96321751-96402024-04-0118338139210.1049/cvi2.12255Point cloud semantic segmentation based on local feature fusion and multilayer attention networkJunjie Wen0Jie Ma1Yuehua Zhao2Tong Nie3Mengxuan Sun4Ziming Fan5School of Electronics and Information Engineering Hebei University of Technology Tianjin ChinaSchool of Electronics and Information Engineering Hebei University of Technology Tianjin ChinaSchool of Electronics and Information Engineering Hebei University of Technology Tianjin ChinaSchool of Electronics and Information Engineering Hebei University of Technology Tianjin ChinaSchool of Electronics and Information Engineering Hebei University of Technology Tianjin ChinaSchool of Electronics and Information Engineering Hebei University of Technology Tianjin ChinaAbstract Semantic segmentation from a three‐dimensional point cloud is vital in autonomous driving, computer vision, and augmented reality. However, current semantic segmentation does not effectively use the point cloud's local geometric features and contextual information, essential for improving segmentation accuracy. A semantic segmentation network that uses local feature fusion and a multilayer attention mechanism is proposed to address these challenges. Specifically, the authors designed a local feature fusion module to encode the geometric and feature information separately, which fully leverages the point cloud's feature perception and geometric structure representation. Furthermore, the authors designed a multilayer attention pooling module consisting of local attention pooling and cascade attention pooling to extract contextual information. Local attention pooling is used to learn local neighbourhood information, and cascade attention pooling captures contextual information from deeper local neighbourhoods. Finally, an enhanced feature representation of important information is obtained by aggregating the features from the two deep attention pooling methods. Extensive experiments on large‐scale point‐cloud datasets Stanford 3D large‐scale indoor spaces and SemanticKITTI indicate that authors network shows excellent advantages over existing representative methods regarding local geometric feature description and global contextual relationships.https://doi.org/10.1049/cvi2.12255computer visionimage segmentationpattern recognition |
| spellingShingle | Junjie Wen Jie Ma Yuehua Zhao Tong Nie Mengxuan Sun Ziming Fan Point cloud semantic segmentation based on local feature fusion and multilayer attention network IET Computer Vision computer vision image segmentation pattern recognition |
| title | Point cloud semantic segmentation based on local feature fusion and multilayer attention network |
| title_full | Point cloud semantic segmentation based on local feature fusion and multilayer attention network |
| title_fullStr | Point cloud semantic segmentation based on local feature fusion and multilayer attention network |
| title_full_unstemmed | Point cloud semantic segmentation based on local feature fusion and multilayer attention network |
| title_short | Point cloud semantic segmentation based on local feature fusion and multilayer attention network |
| title_sort | point cloud semantic segmentation based on local feature fusion and multilayer attention network |
| topic | computer vision image segmentation pattern recognition |
| url | https://doi.org/10.1049/cvi2.12255 |
| work_keys_str_mv | AT junjiewen pointcloudsemanticsegmentationbasedonlocalfeaturefusionandmultilayerattentionnetwork AT jiema pointcloudsemanticsegmentationbasedonlocalfeaturefusionandmultilayerattentionnetwork AT yuehuazhao pointcloudsemanticsegmentationbasedonlocalfeaturefusionandmultilayerattentionnetwork AT tongnie pointcloudsemanticsegmentationbasedonlocalfeaturefusionandmultilayerattentionnetwork AT mengxuansun pointcloudsemanticsegmentationbasedonlocalfeaturefusionandmultilayerattentionnetwork AT zimingfan pointcloudsemanticsegmentationbasedonlocalfeaturefusionandmultilayerattentionnetwork |