Enhanced Self-Attention Network for Remote Sensing Building Change Detection
The self-attention mechanism can break the limitation of the receptive field, model in a global scope, and extract global information efficiently. In this work, we propose a lightweight remote sensing building change detection model (ESACD). In the encoder, we use the enhanced self-attention layer,...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10135088/ |
| _version_ | 1797809189898158080 |
|---|---|
| author | Shike Liang Zhen Hua Jinjiang Li |
| author_facet | Shike Liang Zhen Hua Jinjiang Li |
| author_sort | Shike Liang |
| collection | DOAJ |
| description | The self-attention mechanism can break the limitation of the receptive field, model in a global scope, and extract global information efficiently. In this work, we propose a lightweight remote sensing building change detection model (ESACD). In the encoder, we use the enhanced self-attention layer, CoT layer, instead of the normal convolution operation. The CoT layer fuses the dynamic context with the static context. Compared with the ordinary convolutional layer, this strategy can fully mine the local features between the input keys to dynamically enhance the feature representation. Subsequently, we use dual attention to further mine the low-frequency information and high-frequency information of the images and the semantic features of interest to the model. Dual attention consists of the HiLo attention mechanism and the Tokenizer attention mechanism. HiLo extracts high-frequency information and low-frequency information through two branches. In the high-frequency branch, nonoverlapping windows are applied to the features for self-attention. In the low-frequency branch, average pooling is first performed on features before self-attention. After Tokenizer attention extracts the feature tokens that the model is interested in, it encodes its information and, then, converts the tokens into pixel-level features. Tokenizer attention realizes the efficient extraction of features and enhances the representation ability of the model. Finally, we fuse feature information to enhance the fluidity of information and improve accuracy. Through our experiments on two change detection datasets, ESACD has better performance than other state-of-the-art methods while maintaining fewer parameters. |
| first_indexed | 2024-03-13T06:48:48Z |
| format | Article |
| id | doaj.art-c70a578514b44b07acdeb191b5e2b29c |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-03-13T06:48:48Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-c70a578514b44b07acdeb191b5e2b29c2023-06-07T23:00:14ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352023-01-01164900491510.1109/JSTARS.2023.327872610135088Enhanced Self-Attention Network for Remote Sensing Building Change DetectionShike Liang0Zhen Hua1https://orcid.org/0000-0003-1638-2974Jinjiang Li2https://orcid.org/0000-0002-2080-8678School of Information and Electronic Engineering, Institute of Network Technology, Shandong Technology and Business University, Yantai, ChinaSchool of Computer Science and Technology, Shandong Technology and Business University, Yantai, ChinaSchool of Information and Electronic Engineering, Institute of Network Technology, Shandong Technology and Business University, Yantai, ChinaThe self-attention mechanism can break the limitation of the receptive field, model in a global scope, and extract global information efficiently. In this work, we propose a lightweight remote sensing building change detection model (ESACD). In the encoder, we use the enhanced self-attention layer, CoT layer, instead of the normal convolution operation. The CoT layer fuses the dynamic context with the static context. Compared with the ordinary convolutional layer, this strategy can fully mine the local features between the input keys to dynamically enhance the feature representation. Subsequently, we use dual attention to further mine the low-frequency information and high-frequency information of the images and the semantic features of interest to the model. Dual attention consists of the HiLo attention mechanism and the Tokenizer attention mechanism. HiLo extracts high-frequency information and low-frequency information through two branches. In the high-frequency branch, nonoverlapping windows are applied to the features for self-attention. In the low-frequency branch, average pooling is first performed on features before self-attention. After Tokenizer attention extracts the feature tokens that the model is interested in, it encodes its information and, then, converts the tokens into pixel-level features. Tokenizer attention realizes the efficient extraction of features and enhances the representation ability of the model. Finally, we fuse feature information to enhance the fluidity of information and improve accuracy. Through our experiments on two change detection datasets, ESACD has better performance than other state-of-the-art methods while maintaining fewer parameters.https://ieeexplore.ieee.org/document/10135088/Change detection (CD)remote sensing building imagesself-attention |
| spellingShingle | Shike Liang Zhen Hua Jinjiang Li Enhanced Self-Attention Network for Remote Sensing Building Change Detection IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Change detection (CD) remote sensing building images self-attention |
| title | Enhanced Self-Attention Network for Remote Sensing Building Change Detection |
| title_full | Enhanced Self-Attention Network for Remote Sensing Building Change Detection |
| title_fullStr | Enhanced Self-Attention Network for Remote Sensing Building Change Detection |
| title_full_unstemmed | Enhanced Self-Attention Network for Remote Sensing Building Change Detection |
| title_short | Enhanced Self-Attention Network for Remote Sensing Building Change Detection |
| title_sort | enhanced self attention network for remote sensing building change detection |
| topic | Change detection (CD) remote sensing building images self-attention |
| url | https://ieeexplore.ieee.org/document/10135088/ |
| work_keys_str_mv | AT shikeliang enhancedselfattentionnetworkforremotesensingbuildingchangedetection AT zhenhua enhancedselfattentionnetworkforremotesensingbuildingchangedetection AT jinjiangli enhancedselfattentionnetworkforremotesensingbuildingchangedetection |