Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing Images
Road extraction from optical remote sensing images has drawn much attention in recent decades and has a wide range of applications. Most of the previous studies rarely take into account the unique topological characteristics of the road. It is the most apparent feature of linear structure that descr...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-09-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/12/19/3175 |
| _version_ | 1827705243063812096 |
|---|---|
| author | Kai Geng Xian Sun Zhiyuan Yan Wenhui Diao Xin Gao |
| author_facet | Kai Geng Xian Sun Zhiyuan Yan Wenhui Diao Xin Gao |
| author_sort | Kai Geng |
| collection | DOAJ |
| description | Road extraction from optical remote sensing images has drawn much attention in recent decades and has a wide range of applications. Most of the previous studies rarely take into account the unique topological characteristics of the road. It is the most apparent feature of linear structure that describes the variety of connection relationships of the road. However, designing a particular topological feature extraction network usually results in a model that is too heavy and impractical. To address the problems mentioned above, in this paper, we propose a lightweight topological space network for road extraction based on knowledge distillation (TSKD-Road). Specifically, (1) narrow and short roads easily influence topological features extracted directly in optical remote sensing images. Therefore, we propose a denser teacher network for extracting road structures; (2) to enhance the weight of topological features, we propose a topological space loss calculation model with multiple widths and depths; (3) based on the above innovations, a topological space knowledge distillation framework is proposed, which aims to transfer different kinds of knowledge acquired in a heavy net to a lightweight net, while significantly improving the lightweight net’s accuracy. Experiments were conducted on two publicly available benchmark datasets, which show the obvious superiority and effectiveness of our network. |
| first_indexed | 2024-03-10T15:59:28Z |
| format | Article |
| id | doaj.art-1507656a12c74e59b3ccab6b6e6bfd98 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T15:59:28Z |
| publishDate | 2020-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-1507656a12c74e59b3ccab6b6e6bfd982023-11-20T15:22:57ZengMDPI AGRemote Sensing2072-42922020-09-011219317510.3390/rs12193175Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing ImagesKai Geng0Xian Sun1Zhiyuan Yan2Wenhui Diao3Xin Gao4Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaRoad extraction from optical remote sensing images has drawn much attention in recent decades and has a wide range of applications. Most of the previous studies rarely take into account the unique topological characteristics of the road. It is the most apparent feature of linear structure that describes the variety of connection relationships of the road. However, designing a particular topological feature extraction network usually results in a model that is too heavy and impractical. To address the problems mentioned above, in this paper, we propose a lightweight topological space network for road extraction based on knowledge distillation (TSKD-Road). Specifically, (1) narrow and short roads easily influence topological features extracted directly in optical remote sensing images. Therefore, we propose a denser teacher network for extracting road structures; (2) to enhance the weight of topological features, we propose a topological space loss calculation model with multiple widths and depths; (3) based on the above innovations, a topological space knowledge distillation framework is proposed, which aims to transfer different kinds of knowledge acquired in a heavy net to a lightweight net, while significantly improving the lightweight net’s accuracy. Experiments were conducted on two publicly available benchmark datasets, which show the obvious superiority and effectiveness of our network.https://www.mdpi.com/2072-4292/12/19/3175road extractionlightweight networkknowledge distillationoptical remote sensing imagerytopological structures |
| spellingShingle | Kai Geng Xian Sun Zhiyuan Yan Wenhui Diao Xin Gao Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing Images Remote Sensing road extraction lightweight network knowledge distillation optical remote sensing imagery topological structures |
| title | Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing Images |
| title_full | Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing Images |
| title_fullStr | Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing Images |
| title_full_unstemmed | Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing Images |
| title_short | Topological Space Knowledge Distillation for Compact Road Extraction in Optical Remote Sensing Images |
| title_sort | topological space knowledge distillation for compact road extraction in optical remote sensing images |
| topic | road extraction lightweight network knowledge distillation optical remote sensing imagery topological structures |
| url | https://www.mdpi.com/2072-4292/12/19/3175 |
| work_keys_str_mv | AT kaigeng topologicalspaceknowledgedistillationforcompactroadextractioninopticalremotesensingimages AT xiansun topologicalspaceknowledgedistillationforcompactroadextractioninopticalremotesensingimages AT zhiyuanyan topologicalspaceknowledgedistillationforcompactroadextractioninopticalremotesensingimages AT wenhuidiao topologicalspaceknowledgedistillationforcompactroadextractioninopticalremotesensingimages AT xingao topologicalspaceknowledgedistillationforcompactroadextractioninopticalremotesensingimages |