Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification
Abstract An occluded person re‐identification (ReID) approach is presented by constructing a Bi‐level deep Mutual learning assisted Multi‐task network (BMM), where the holistic and occluded person ReID tasks are treated as two related but not identical tasks. This is inspired by the human perception...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2023-03-01
|
Series: | IET Image Processing |
Online Access: | https://doi.org/10.1049/ipr2.12688 |
_version_ | 1811159626970824704 |
---|---|
author | Yi Wang Liangbo Wang Yu Zhou |
author_facet | Yi Wang Liangbo Wang Yu Zhou |
author_sort | Yi Wang |
collection | DOAJ |
description | Abstract An occluded person re‐identification (ReID) approach is presented by constructing a Bi‐level deep Mutual learning assisted Multi‐task network (BMM), where the holistic and occluded person ReID tasks are treated as two related but not identical tasks. This is inspired by the human perception characteristic that there exist both similarities and differences when human views a holistic image and the occluded one. Specifically, a multi‐task network with two branches is designed, where the convolutional neural network based feature representation part shares the weights by two tasks for commonality extraction, while the following output layers have respective weights for difference representation. Furthermore, as the non‐occluded regions convey discriminative information, a bi‐level mutual learning strategy is proposed and applied mutually on two branches to obtain more effective information from the non‐occluded regions in the occluded images for better identity recognition. This is achieved by both feature‐level and output‐level mutual loss functions. Extensive experiments prove the advantages of the BMM for person ReID. |
first_indexed | 2024-04-10T05:44:20Z |
format | Article |
id | doaj.art-7eb43d96ef7143cf89aedf70a6aedafd |
institution | Directory Open Access Journal |
issn | 1751-9659 1751-9667 |
language | English |
last_indexed | 2024-04-10T05:44:20Z |
publishDate | 2023-03-01 |
publisher | Wiley |
record_format | Article |
series | IET Image Processing |
spelling | doaj.art-7eb43d96ef7143cf89aedf70a6aedafd2023-03-06T04:27:52ZengWileyIET Image Processing1751-96591751-96672023-03-0117497998710.1049/ipr2.12688Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identificationYi Wang0Liangbo Wang1Yu Zhou2Jiangsu Normal University Kewen College Xuzhou ChinaSchool of Information and Control Engineering China University of Mining and Technology Xuzhou ChinaSchool of Information and Control Engineering China University of Mining and Technology Xuzhou ChinaAbstract An occluded person re‐identification (ReID) approach is presented by constructing a Bi‐level deep Mutual learning assisted Multi‐task network (BMM), where the holistic and occluded person ReID tasks are treated as two related but not identical tasks. This is inspired by the human perception characteristic that there exist both similarities and differences when human views a holistic image and the occluded one. Specifically, a multi‐task network with two branches is designed, where the convolutional neural network based feature representation part shares the weights by two tasks for commonality extraction, while the following output layers have respective weights for difference representation. Furthermore, as the non‐occluded regions convey discriminative information, a bi‐level mutual learning strategy is proposed and applied mutually on two branches to obtain more effective information from the non‐occluded regions in the occluded images for better identity recognition. This is achieved by both feature‐level and output‐level mutual loss functions. Extensive experiments prove the advantages of the BMM for person ReID.https://doi.org/10.1049/ipr2.12688 |
spellingShingle | Yi Wang Liangbo Wang Yu Zhou Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification IET Image Processing |
title | Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification |
title_full | Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification |
title_fullStr | Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification |
title_full_unstemmed | Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification |
title_short | Bi‐level deep mutual learning assisted multi‐task network for occluded person re‐identification |
title_sort | bi level deep mutual learning assisted multi task network for occluded person re identification |
url | https://doi.org/10.1049/ipr2.12688 |
work_keys_str_mv | AT yiwang bileveldeepmutuallearningassistedmultitasknetworkforoccludedpersonreidentification AT liangbowang bileveldeepmutuallearningassistedmultitasknetworkforoccludedpersonreidentification AT yuzhou bileveldeepmutuallearningassistedmultitasknetworkforoccludedpersonreidentification |