A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning
To address the issues of high computational complexity and insufficient aggregation of global and local information in existing image segmentation methods, this paper proposes an efficient segmentation model based on Spatial Context Learning, named SCLSeg. The main idea is to aggregate local regions...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10759633/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841533424380149760 |
|---|---|
| author | Xiaomei Xiao Jialiang Tang Xiaoyan Lu Zhengyong Feng Yi Li |
| author_facet | Xiaomei Xiao Jialiang Tang Xiaoyan Lu Zhengyong Feng Yi Li |
| author_sort | Xiaomei Xiao |
| collection | DOAJ |
| description | To address the issues of high computational complexity and insufficient aggregation of global and local information in existing image segmentation methods, this paper proposes an efficient segmentation model based on Spatial Context Learning, named SCLSeg. The main idea is to aggregate local regions into higher-level semantic regions in a learnable manner. The proposed Spatial Context Guided Feature Alignment module (SC-FA) learns aligned features from image-level to local regions, exploring and integrating contextual information. During training, a multi-scale strategy is used to group semantic regions, and a Channel Aggregation Block (CAB) is designed to dynamically capture semantic groups through a mechanism of feature separation and fusion, thereby aggregating multi-level pixel features to generate the final segmentation results. We further introduce a boundary loss to optimize the accuracy of segmentation edges. To meet real-time processing requirements, a series of lightweight strategies and simplified structures are adopted to reduce computational costs, including lightweight encoding, channel compression, and simplified neck. Our method achieves good performance on the Cityscapes and Camvid datasets, specifically achieving 76.45% mIoU & 237 FPS on the Cityscapes test set, and 73.95% mIoU & 300.4 FPS on the CamVid test set. |
| format | Article |
| id | doaj-art-935feab7933a4dc6ad58840959beb071 |
| institution | Kabale University |
| issn | 2169-3536 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-935feab7933a4dc6ad58840959beb0712025-01-16T00:01:37ZengIEEEIEEE Access2169-35362024-01-011217849517850610.1109/ACCESS.2024.350367610759633A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context LearningXiaomei Xiao0https://orcid.org/0009-0000-0326-4448Jialiang Tang1https://orcid.org/0009-0003-1387-1403Xiaoyan Lu2Zhengyong Feng3https://orcid.org/0009-0005-7291-1684Yi Li4School of Electronic Information Engineering, Electronic Information Processing Engineering Technology Research Center, China West Normal University, Nanchong, ChinaSchool of Electronic Information Engineering, Electronic Information Processing Engineering Technology Research Center, China West Normal University, Nanchong, ChinaSchool of Electronic Information Engineering, Electronic Information Processing Engineering Technology Research Center, China West Normal University, Nanchong, ChinaSchool of Electronic Information Engineering, Electronic Information Processing Engineering Technology Research Center, China West Normal University, Nanchong, ChinaCollege of Physics and Engineering Technology, Chengdu Normal University, Chengdu, ChinaTo address the issues of high computational complexity and insufficient aggregation of global and local information in existing image segmentation methods, this paper proposes an efficient segmentation model based on Spatial Context Learning, named SCLSeg. The main idea is to aggregate local regions into higher-level semantic regions in a learnable manner. The proposed Spatial Context Guided Feature Alignment module (SC-FA) learns aligned features from image-level to local regions, exploring and integrating contextual information. During training, a multi-scale strategy is used to group semantic regions, and a Channel Aggregation Block (CAB) is designed to dynamically capture semantic groups through a mechanism of feature separation and fusion, thereby aggregating multi-level pixel features to generate the final segmentation results. We further introduce a boundary loss to optimize the accuracy of segmentation edges. To meet real-time processing requirements, a series of lightweight strategies and simplified structures are adopted to reduce computational costs, including lightweight encoding, channel compression, and simplified neck. Our method achieves good performance on the Cityscapes and Camvid datasets, specifically achieving 76.45% mIoU & 237 FPS on the Cityscapes test set, and 73.95% mIoU & 300.4 FPS on the CamVid test set.https://ieeexplore.ieee.org/document/10759633/Real-time semantic segmentationspatial context guidancefeature attentionfeature alignment |
| spellingShingle | Xiaomei Xiao Jialiang Tang Xiaoyan Lu Zhengyong Feng Yi Li A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning IEEE Access Real-time semantic segmentation spatial context guidance feature attention feature alignment |
| title | A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning |
| title_full | A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning |
| title_fullStr | A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning |
| title_full_unstemmed | A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning |
| title_short | A Real-Time Road Scene Semantic Segmentation Model Based on Spatial Context Learning |
| title_sort | real time road scene semantic segmentation model based on spatial context learning |
| topic | Real-time semantic segmentation spatial context guidance feature attention feature alignment |
| url | https://ieeexplore.ieee.org/document/10759633/ |
| work_keys_str_mv | AT xiaomeixiao arealtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT jialiangtang arealtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT xiaoyanlu arealtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT zhengyongfeng arealtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT yili arealtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT xiaomeixiao realtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT jialiangtang realtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT xiaoyanlu realtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT zhengyongfeng realtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning AT yili realtimeroadscenesemanticsegmentationmodelbasedonspatialcontextlearning |