Affinity Learning Via Self-Supervised Diffusion for Spectral Clustering
Spectral clustering makes use of the spectrum of an input affinity matrix to segment data into disjoint clusters. The performance of spectral clustering depends heavily on the quality of the affinity matrix. Commonly used affinity matrices are constructed by either the Gaussian kernel or the self-ex...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9293293/ |
| _version_ | 1818921927158267904 |
|---|---|
| author | Jianfeng Ye Qilin Li Jinlong Yu Xincheng Wang Huaming Wang |
| author_facet | Jianfeng Ye Qilin Li Jinlong Yu Xincheng Wang Huaming Wang |
| author_sort | Jianfeng Ye |
| collection | DOAJ |
| description | Spectral clustering makes use of the spectrum of an input affinity matrix to segment data into disjoint clusters. The performance of spectral clustering depends heavily on the quality of the affinity matrix. Commonly used affinity matrices are constructed by either the Gaussian kernel or the self-expressive model with sparse or low-rank constraints. A technique called diffusion which acts as a post-process has recently shown to improve the quality of the affinity matrix significantly, by taking advantage of the contextual information. In this paper, we propose a variant of the diffusion process, named Self-Supervised Diffusion, which incorporates clustering result as feedback to provide supervisory signals for the diffusion process. The proposed method contains two stages, namely affinity learning with diffusion and spectral clustering. It works in an iterative fashion, where in each iteration the clustering result is utilized to calculate a pseudo-label similarity so that it can aid the affinity learning stage in the next iteration. Extensive experiments on both synthetic and real-world data have demonstrated that the proposed method can learn accurate and robust affinity, and thus achieves superior clustering performance. |
| first_indexed | 2024-12-20T01:45:25Z |
| format | Article |
| id | doaj.art-e637f02f4bc448d0b6c0143ef9a7e404 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-20T01:45:25Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-e637f02f4bc448d0b6c0143ef9a7e4042022-12-21T19:57:48ZengIEEEIEEE Access2169-35362021-01-0197170718210.1109/ACCESS.2020.30446969293293Affinity Learning Via Self-Supervised Diffusion for Spectral ClusteringJianfeng Ye0https://orcid.org/0000-0003-3100-5587Qilin Li1https://orcid.org/0000-0001-6584-8879Jinlong Yu2Xincheng Wang3Huaming Wang4https://orcid.org/0000-0002-4434-7482Department of Electromechanical, Nanjing University of Aeronautics and Astronautics, Nanjing, ChinaDepartment of Computing, Curtin University, Perth, WA, AustraliaDepartment of Electromechanical, Nanjing University of Aeronautics and Astronautics, Nanjing, ChinaDepartment of Electromechanical, Nanjing University of Aeronautics and Astronautics, Nanjing, ChinaDepartment of Electromechanical, Nanjing University of Aeronautics and Astronautics, Nanjing, ChinaSpectral clustering makes use of the spectrum of an input affinity matrix to segment data into disjoint clusters. The performance of spectral clustering depends heavily on the quality of the affinity matrix. Commonly used affinity matrices are constructed by either the Gaussian kernel or the self-expressive model with sparse or low-rank constraints. A technique called diffusion which acts as a post-process has recently shown to improve the quality of the affinity matrix significantly, by taking advantage of the contextual information. In this paper, we propose a variant of the diffusion process, named Self-Supervised Diffusion, which incorporates clustering result as feedback to provide supervisory signals for the diffusion process. The proposed method contains two stages, namely affinity learning with diffusion and spectral clustering. It works in an iterative fashion, where in each iteration the clustering result is utilized to calculate a pseudo-label similarity so that it can aid the affinity learning stage in the next iteration. Extensive experiments on both synthetic and real-world data have demonstrated that the proposed method can learn accurate and robust affinity, and thus achieves superior clustering performance.https://ieeexplore.ieee.org/document/9293293/Affinity learningdiffusion processspectral clustering |
| spellingShingle | Jianfeng Ye Qilin Li Jinlong Yu Xincheng Wang Huaming Wang Affinity Learning Via Self-Supervised Diffusion for Spectral Clustering IEEE Access Affinity learning diffusion process spectral clustering |
| title | Affinity Learning Via Self-Supervised Diffusion for Spectral Clustering |
| title_full | Affinity Learning Via Self-Supervised Diffusion for Spectral Clustering |
| title_fullStr | Affinity Learning Via Self-Supervised Diffusion for Spectral Clustering |
| title_full_unstemmed | Affinity Learning Via Self-Supervised Diffusion for Spectral Clustering |
| title_short | Affinity Learning Via Self-Supervised Diffusion for Spectral Clustering |
| title_sort | affinity learning via self supervised diffusion for spectral clustering |
| topic | Affinity learning diffusion process spectral clustering |
| url | https://ieeexplore.ieee.org/document/9293293/ |
| work_keys_str_mv | AT jianfengye affinitylearningviaselfsuperviseddiffusionforspectralclustering AT qilinli affinitylearningviaselfsuperviseddiffusionforspectralclustering AT jinlongyu affinitylearningviaselfsuperviseddiffusionforspectralclustering AT xinchengwang affinitylearningviaselfsuperviseddiffusionforspectralclustering AT huamingwang affinitylearningviaselfsuperviseddiffusionforspectralclustering |