AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral Videos
Object tracking using Hyperspectral Images (HSIs) obtains satisfactory result in distinguishing objects with similar colors. Yet, the tracking algorithm tends to fail when the target undergoes deformation. In this paper, a SiamRPN based hyperspectral tracker is proposed to deal with this problem. Fi...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-03-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/15/7/1731 |
| _version_ | 1797607114206609408 |
|---|---|
| author | Shiqing Wang Kun Qian Jianlu Shen Hongyu Ma Peng Chen |
| author_facet | Shiqing Wang Kun Qian Jianlu Shen Hongyu Ma Peng Chen |
| author_sort | Shiqing Wang |
| collection | DOAJ |
| description | Object tracking using Hyperspectral Images (HSIs) obtains satisfactory result in distinguishing objects with similar colors. Yet, the tracking algorithm tends to fail when the target undergoes deformation. In this paper, a SiamRPN based hyperspectral tracker is proposed to deal with this problem. Firstly, a band selection method based on a genetic optimization method is designed for rapidly reducing the redundancy of information in HSIs. Specifically, three bands with highest joint entropy are selected. To solve the problem that the information of the template in the SiamRPN model decays over time, an update network is trained on the dataset from general objective tracking benchmark, which can obtain effective cumulative templates. The use of cumulative templates with spectral information makes it easier to track the deformed target. In addition, transfer learning of the pre-trained SiamRPN is designed to obtain a better model for HSIs. The experimental results show that the proposed tracker can obtain good tracking results over the entire public dataset, and that it is better than the other popular trackers when the target’s deformation is qualitatively and quantitatively compared, achieving an overall success rate of 57.5% and a deformation challenge success rate of 70.8%. |
| first_indexed | 2024-03-11T05:26:39Z |
| format | Article |
| id | doaj.art-5372b3cc6d14423f94fca66490b983cf |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-11T05:26:39Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-5372b3cc6d14423f94fca66490b983cf2023-11-17T17:28:05ZengMDPI AGRemote Sensing2072-42922023-03-01157173110.3390/rs15071731AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral VideosShiqing Wang0Kun Qian1Jianlu Shen2Hongyu Ma3Peng Chen4School of Artifical Intelligence and Computer Science, Jiangnan University, Wuxi 214122, ChinaSchool of Artifical Intelligence and Computer Science, Jiangnan University, Wuxi 214122, ChinaSchool of Artifical Intelligence and Computer Science, Jiangnan University, Wuxi 214122, ChinaCollege of Automation, Wuxi University, Wuxi 214122, ChinaSchool of Artifical Intelligence and Computer Science, Jiangnan University, Wuxi 214122, ChinaObject tracking using Hyperspectral Images (HSIs) obtains satisfactory result in distinguishing objects with similar colors. Yet, the tracking algorithm tends to fail when the target undergoes deformation. In this paper, a SiamRPN based hyperspectral tracker is proposed to deal with this problem. Firstly, a band selection method based on a genetic optimization method is designed for rapidly reducing the redundancy of information in HSIs. Specifically, three bands with highest joint entropy are selected. To solve the problem that the information of the template in the SiamRPN model decays over time, an update network is trained on the dataset from general objective tracking benchmark, which can obtain effective cumulative templates. The use of cumulative templates with spectral information makes it easier to track the deformed target. In addition, transfer learning of the pre-trained SiamRPN is designed to obtain a better model for HSIs. The experimental results show that the proposed tracker can obtain good tracking results over the entire public dataset, and that it is better than the other popular trackers when the target’s deformation is qualitatively and quantitatively compared, achieving an overall success rate of 57.5% and a deformation challenge success rate of 70.8%.https://www.mdpi.com/2072-4292/15/7/1731object trackinghyperspectral imagessiameseintelligent optimizationanti-deformation |
| spellingShingle | Shiqing Wang Kun Qian Jianlu Shen Hongyu Ma Peng Chen AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral Videos Remote Sensing object tracking hyperspectral images siamese intelligent optimization anti-deformation |
| title | AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral Videos |
| title_full | AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral Videos |
| title_fullStr | AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral Videos |
| title_full_unstemmed | AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral Videos |
| title_short | AD-SiamRPN: Anti-Deformation Object Tracking via an Improved Siamese Region Proposal Network on Hyperspectral Videos |
| title_sort | ad siamrpn anti deformation object tracking via an improved siamese region proposal network on hyperspectral videos |
| topic | object tracking hyperspectral images siamese intelligent optimization anti-deformation |
| url | https://www.mdpi.com/2072-4292/15/7/1731 |
| work_keys_str_mv | AT shiqingwang adsiamrpnantideformationobjecttrackingviaanimprovedsiameseregionproposalnetworkonhyperspectralvideos AT kunqian adsiamrpnantideformationobjecttrackingviaanimprovedsiameseregionproposalnetworkonhyperspectralvideos AT jianlushen adsiamrpnantideformationobjecttrackingviaanimprovedsiameseregionproposalnetworkonhyperspectralvideos AT hongyuma adsiamrpnantideformationobjecttrackingviaanimprovedsiameseregionproposalnetworkonhyperspectralvideos AT pengchen adsiamrpnantideformationobjecttrackingviaanimprovedsiameseregionproposalnetworkonhyperspectralvideos |