Person re-identification for video surveillance
Person Re-Identification is a new technique compared to the person identification, which is more difficult and complex. It is easily affected by the surrounding environment and the person's gestures. Although Person Re-ID has made great progress in recent years, there are still a lot of problem...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2017
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/72552 |
| _version_ | 1811681201539252224 |
|---|---|
| author | Tang, Jinli |
| author2 | Chau Lap Pui |
| author_facet | Chau Lap Pui Tang, Jinli |
| author_sort | Tang, Jinli |
| collection | NTU |
| description | Person Re-Identification is a new technique compared to the person identification, which is more difficult and complex. It is easily affected by the surrounding environment and the person's gestures. Although Person Re-ID has made great progress in recent years, there are still a lot of problems that need to be solved. In this thesis, the basic knowledge about person identification is introduced and reviewed. It then we introduced the image based Person Re-ID and video based Person Re-ID. For the theory and method, the TDL distance learning model was studied. Besides the TDL based method, the methods involving feature extration are also introduced. For the color feature, we used the localised average colour histogram. For the space-time feature, the HOG3d descriptor was used. Two datasets including iLIDS and PRID2011 were used for the experiment. Both experiments obtained good results: the Rank 1 rate for iLIDS is about 57% and the Rank 1 rate for PRID2011 is about 55%. They are better than the previous results, which lays the foundation for future work. |
| first_indexed | 2024-10-01T03:37:11Z |
| format | Thesis |
| id | ntu-10356/72552 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:37:11Z |
| publishDate | 2017 |
| record_format | dspace |
| spelling | ntu-10356/725522023-07-04T15:53:50Z Person re-identification for video surveillance Tang, Jinli Chau Lap Pui School of Electrical and Electronic Engineering DRNTU::Engineering::Electrical and electronic engineering Person Re-Identification is a new technique compared to the person identification, which is more difficult and complex. It is easily affected by the surrounding environment and the person's gestures. Although Person Re-ID has made great progress in recent years, there are still a lot of problems that need to be solved. In this thesis, the basic knowledge about person identification is introduced and reviewed. It then we introduced the image based Person Re-ID and video based Person Re-ID. For the theory and method, the TDL distance learning model was studied. Besides the TDL based method, the methods involving feature extration are also introduced. For the color feature, we used the localised average colour histogram. For the space-time feature, the HOG3d descriptor was used. Two datasets including iLIDS and PRID2011 were used for the experiment. Both experiments obtained good results: the Rank 1 rate for iLIDS is about 57% and the Rank 1 rate for PRID2011 is about 55%. They are better than the previous results, which lays the foundation for future work. Master of Science (Communications Engineering) 2017-08-28T11:35:34Z 2017-08-28T11:35:34Z 2017 Thesis http://hdl.handle.net/10356/72552 en 57 p. application/pdf |
| spellingShingle | DRNTU::Engineering::Electrical and electronic engineering Tang, Jinli Person re-identification for video surveillance |
| title | Person re-identification for video surveillance |
| title_full | Person re-identification for video surveillance |
| title_fullStr | Person re-identification for video surveillance |
| title_full_unstemmed | Person re-identification for video surveillance |
| title_short | Person re-identification for video surveillance |
| title_sort | person re identification for video surveillance |
| topic | DRNTU::Engineering::Electrical and electronic engineering |
| url | http://hdl.handle.net/10356/72552 |
| work_keys_str_mv | AT tangjinli personreidentificationforvideosurveillance |