HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image Segmentation
The model, Transformer, is known to rely on a self-attention mechanism to model distant dependencies, which focuses on modeling the dependencies of the global elements. However, its sensitivity to the local details of the foreground information is not significant. Local detail features help to ident...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-09-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/22/18/7024 |
| _version_ | 1827656764014723072 |
|---|---|
| author | Lijing An Liejun Wang Yongming Li |
| author_facet | Lijing An Liejun Wang Yongming Li |
| author_sort | Lijing An |
| collection | DOAJ |
| description | The model, Transformer, is known to rely on a self-attention mechanism to model distant dependencies, which focuses on modeling the dependencies of the global elements. However, its sensitivity to the local details of the foreground information is not significant. Local detail features help to identify the blurred boundaries in medical images more accurately. In order to make up for the defects of Transformer and capture more abundant local information, this paper proposes an attention and MLP hybrid-encoder architecture combining the Efficient Attention Module (EAM) with a Dual-channel Shift MLP module (DS-MLP), called HEA-Net. Specifically, we effectively connect the convolution block with Transformer through EAM to enhance the foreground and suppress the invalid background information in medical images. Meanwhile, DS-MLP further enhances the foreground information via channel and spatial shift operations. Extensive experiments on public datasets confirm the excellent performance of our proposed HEA-Net. In particular, on the GlaS and MoNuSeg datasets, the Dice reached 90.56% and 80.80%, respectively, and the IoU reached 83.62% and 68.26%, respectively. |
| first_indexed | 2024-03-09T22:32:56Z |
| format | Article |
| id | doaj.art-7721c327c43849f7959cd9df9ee06e94 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-09T22:32:56Z |
| publishDate | 2022-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-7721c327c43849f7959cd9df9ee06e942023-11-23T18:53:14ZengMDPI AGSensors1424-82202022-09-012218702410.3390/s22187024HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image SegmentationLijing An0Liejun Wang1Yongming Li2College of Information Science and Engineering, Xinjiang University, Urumqi 830000, ChinaCollege of Information Science and Engineering, Xinjiang University, Urumqi 830000, ChinaCollege of Information Science and Engineering, Xinjiang University, Urumqi 830000, ChinaThe model, Transformer, is known to rely on a self-attention mechanism to model distant dependencies, which focuses on modeling the dependencies of the global elements. However, its sensitivity to the local details of the foreground information is not significant. Local detail features help to identify the blurred boundaries in medical images more accurately. In order to make up for the defects of Transformer and capture more abundant local information, this paper proposes an attention and MLP hybrid-encoder architecture combining the Efficient Attention Module (EAM) with a Dual-channel Shift MLP module (DS-MLP), called HEA-Net. Specifically, we effectively connect the convolution block with Transformer through EAM to enhance the foreground and suppress the invalid background information in medical images. Meanwhile, DS-MLP further enhances the foreground information via channel and spatial shift operations. Extensive experiments on public datasets confirm the excellent performance of our proposed HEA-Net. In particular, on the GlaS and MoNuSeg datasets, the Dice reached 90.56% and 80.80%, respectively, and the IoU reached 83.62% and 68.26%, respectively.https://www.mdpi.com/1424-8220/22/18/7024attentionMLPTransformer |
| spellingShingle | Lijing An Liejun Wang Yongming Li HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image Segmentation Sensors attention MLP Transformer |
| title | HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image Segmentation |
| title_full | HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image Segmentation |
| title_fullStr | HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image Segmentation |
| title_full_unstemmed | HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image Segmentation |
| title_short | HEA-Net: Attention and MLP Hybrid Encoder Architecture for Medical Image Segmentation |
| title_sort | hea net attention and mlp hybrid encoder architecture for medical image segmentation |
| topic | attention MLP Transformer |
| url | https://www.mdpi.com/1424-8220/22/18/7024 |
| work_keys_str_mv | AT lijingan heanetattentionandmlphybridencoderarchitectureformedicalimagesegmentation AT liejunwang heanetattentionandmlphybridencoderarchitectureformedicalimagesegmentation AT yongmingli heanetattentionandmlphybridencoderarchitectureformedicalimagesegmentation |