Exploring New Backbone and Attention Module for Semantic Segmentation in Street Scenes
Semantic segmentation, as dense pixel-wise classification task, played an important tache in scene understanding. There are two main challenges in many state-of-the-art works: 1) most backbone of segmentation models that often were extracted from pretrained classification models generated poor perfo...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2018-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/8531594/ |
| _version_ | 1819133500450668544 |
|---|---|
| author | Lei Fan Wei-Chien Wang Fuyuan Zha Jiapeng Yan |
| author_facet | Lei Fan Wei-Chien Wang Fuyuan Zha Jiapeng Yan |
| author_sort | Lei Fan |
| collection | DOAJ |
| description | Semantic segmentation, as dense pixel-wise classification task, played an important tache in scene understanding. There are two main challenges in many state-of-the-art works: 1) most backbone of segmentation models that often were extracted from pretrained classification models generated poor performance in small categories because they were lacking in spatial information and 2) the gap of combination between high-level and low-level features in segmentation models has led to inaccurate predictions. To handle these challenges, in this paper, we proposed a new tailored backbone and attention select module for segmentation tasks. Specifically, our new backbone was modified from the original Resnet, which can yield better segmentation performance. Attention select module employed spatial and channel self-attention mechanism to reinforce the propagation of contextual features, which can aggregate semantic and spatial information simultaneously. In addition, based on our new backbone and attention select module, we further proposed our segmentation model for street scenes understanding. We conducted a series of ablation studies on two public benchmarks, including Cityscapes and CamVid dataset to demonstrate the effectiveness of our proposals. Our model achieved a mIoU score of 71.5% on the Cityscapes test set with only fine annotation data and 60.1% on the CamVid test set. |
| first_indexed | 2024-12-22T09:48:17Z |
| format | Article |
| id | doaj.art-f8bcf876de0e481183632296649f28ed |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-22T09:48:17Z |
| publishDate | 2018-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-f8bcf876de0e481183632296649f28ed2022-12-21T18:30:29ZengIEEEIEEE Access2169-35362018-01-016715667158010.1109/ACCESS.2018.28808778531594Exploring New Backbone and Attention Module for Semantic Segmentation in Street ScenesLei Fan0https://orcid.org/0000-0001-9472-7152Wei-Chien Wang1Fuyuan Zha2Jiapeng Yan3School of Electrical Engineering and Automation, Hefei University of Technology, Hefei, ChinaScience and Engineering Faculty, Queensland University of Technology, Brisbane, QLD, AustraliaSchool of Electrical Engineering and Automation, Hefei University of Technology, Hefei, ChinaSchool of Electrical Engineering and Automation, Hefei University of Technology, Hefei, ChinaSemantic segmentation, as dense pixel-wise classification task, played an important tache in scene understanding. There are two main challenges in many state-of-the-art works: 1) most backbone of segmentation models that often were extracted from pretrained classification models generated poor performance in small categories because they were lacking in spatial information and 2) the gap of combination between high-level and low-level features in segmentation models has led to inaccurate predictions. To handle these challenges, in this paper, we proposed a new tailored backbone and attention select module for segmentation tasks. Specifically, our new backbone was modified from the original Resnet, which can yield better segmentation performance. Attention select module employed spatial and channel self-attention mechanism to reinforce the propagation of contextual features, which can aggregate semantic and spatial information simultaneously. In addition, based on our new backbone and attention select module, we further proposed our segmentation model for street scenes understanding. We conducted a series of ablation studies on two public benchmarks, including Cityscapes and CamVid dataset to demonstrate the effectiveness of our proposals. Our model achieved a mIoU score of 71.5% on the Cityscapes test set with only fine annotation data and 60.1% on the CamVid test set.https://ieeexplore.ieee.org/document/8531594/Semantic segmentationsegmentation backboneattention mechanismstreet scenes |
| spellingShingle | Lei Fan Wei-Chien Wang Fuyuan Zha Jiapeng Yan Exploring New Backbone and Attention Module for Semantic Segmentation in Street Scenes IEEE Access Semantic segmentation segmentation backbone attention mechanism street scenes |
| title | Exploring New Backbone and Attention Module for Semantic Segmentation in Street Scenes |
| title_full | Exploring New Backbone and Attention Module for Semantic Segmentation in Street Scenes |
| title_fullStr | Exploring New Backbone and Attention Module for Semantic Segmentation in Street Scenes |
| title_full_unstemmed | Exploring New Backbone and Attention Module for Semantic Segmentation in Street Scenes |
| title_short | Exploring New Backbone and Attention Module for Semantic Segmentation in Street Scenes |
| title_sort | exploring new backbone and attention module for semantic segmentation in street scenes |
| topic | Semantic segmentation segmentation backbone attention mechanism street scenes |
| url | https://ieeexplore.ieee.org/document/8531594/ |
| work_keys_str_mv | AT leifan exploringnewbackboneandattentionmoduleforsemanticsegmentationinstreetscenes AT weichienwang exploringnewbackboneandattentionmoduleforsemanticsegmentationinstreetscenes AT fuyuanzha exploringnewbackboneandattentionmoduleforsemanticsegmentationinstreetscenes AT jiapengyan exploringnewbackboneandattentionmoduleforsemanticsegmentationinstreetscenes |