Pedestrian re‐identification via coarse‐to‐fine ranking
Appearance‐based person re‐identification is particularly difficult due to varying lighting conditions and pose variations across camera views. Taking inspiration from image retrieval, in which windowed searching over locations is proven to be more effective, the authors first perform dense local fe...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2016-08-01
|
| Series: | IET Computer Vision |
| Subjects: | |
| Online Access: | https://doi.org/10.1049/iet-cvi.2014.0288 |
| _version_ | 1797684516192518144 |
|---|---|
| author | Liu Xiaokai |
| author_facet | Liu Xiaokai |
| author_sort | Liu Xiaokai |
| collection | DOAJ |
| description | Appearance‐based person re‐identification is particularly difficult due to varying lighting conditions and pose variations across camera views. Taking inspiration from image retrieval, in which windowed searching over locations is proven to be more effective, the authors first perform dense local feature matching using graph cuts to properly deal with the pose variation problem. However, the re‐identification problem suffers from far more overlap between feature distributions. In a re‐identification problem, many samples cropped from surveillance videos are heavily contaminated by external factors or internal mechanical noises, making the images from the same pedestrian totally different. These overly difficult samples would significantly degenerate the training performance. To address this problem, a query‐level loss function for ranking is proposed, benefiting from taking into account the training data every query set to decrease the punishment for those morbid samples. The authors further develop a coarse‐to‐fine iterative algorithm, where the update in each iteration is computed by solving a gradient‐based optimisation and update iteration is to refine the training data by adjusting an ‘Expected rank’ parameter. The authors present experiments to demonstrate the performance gain of the proposed method over existing template matching and ranking models. |
| first_indexed | 2024-03-12T00:30:47Z |
| format | Article |
| id | doaj.art-0370d1c878e64951a69e9b1f3f09013c |
| institution | Directory Open Access Journal |
| issn | 1751-9632 1751-9640 |
| language | English |
| last_indexed | 2024-03-12T00:30:47Z |
| publishDate | 2016-08-01 |
| publisher | Wiley |
| record_format | Article |
| series | IET Computer Vision |
| spelling | doaj.art-0370d1c878e64951a69e9b1f3f09013c2023-09-15T10:20:52ZengWileyIET Computer Vision1751-96321751-96402016-08-0110536837510.1049/iet-cvi.2014.0288Pedestrian re‐identification via coarse‐to‐fine rankingLiu Xiaokai0Faculty of Electronic Information and Electrical EngineeringDalian University of TechnologyDalian116023People's Republic of ChinaAppearance‐based person re‐identification is particularly difficult due to varying lighting conditions and pose variations across camera views. Taking inspiration from image retrieval, in which windowed searching over locations is proven to be more effective, the authors first perform dense local feature matching using graph cuts to properly deal with the pose variation problem. However, the re‐identification problem suffers from far more overlap between feature distributions. In a re‐identification problem, many samples cropped from surveillance videos are heavily contaminated by external factors or internal mechanical noises, making the images from the same pedestrian totally different. These overly difficult samples would significantly degenerate the training performance. To address this problem, a query‐level loss function for ranking is proposed, benefiting from taking into account the training data every query set to decrease the punishment for those morbid samples. The authors further develop a coarse‐to‐fine iterative algorithm, where the update in each iteration is computed by solving a gradient‐based optimisation and update iteration is to refine the training data by adjusting an ‘Expected rank’ parameter. The authors present experiments to demonstrate the performance gain of the proposed method over existing template matching and ranking models.https://doi.org/10.1049/iet-cvi.2014.0288pedestrian re-identificationcoarse-to-fine rankingappearance-based person re-identificationcamera viewsre-identification problemreidentification problem |
| spellingShingle | Liu Xiaokai Pedestrian re‐identification via coarse‐to‐fine ranking IET Computer Vision pedestrian re-identification coarse-to-fine ranking appearance-based person re-identification camera views re-identification problem reidentification problem |
| title | Pedestrian re‐identification via coarse‐to‐fine ranking |
| title_full | Pedestrian re‐identification via coarse‐to‐fine ranking |
| title_fullStr | Pedestrian re‐identification via coarse‐to‐fine ranking |
| title_full_unstemmed | Pedestrian re‐identification via coarse‐to‐fine ranking |
| title_short | Pedestrian re‐identification via coarse‐to‐fine ranking |
| title_sort | pedestrian re identification via coarse to fine ranking |
| topic | pedestrian re-identification coarse-to-fine ranking appearance-based person re-identification camera views re-identification problem reidentification problem |
| url | https://doi.org/10.1049/iet-cvi.2014.0288 |
| work_keys_str_mv | AT liuxiaokai pedestrianreidentificationviacoarsetofineranking |