Real time multiple object tracking using deep features and localization information
In this paper we propose a tracking by detection method using a dissimilarity measure calculated based on the location and the appearance information of the object. These dissimilarity values are used in Hungarian Algorithm [1] in the data association step for track identity assignment. We make use...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Conference Paper |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/141797 |
| _version_ | 1824457126780600320 |
|---|---|
| author | Karunasekera, Hasith Zhang, Handuo Wang, Han |
| author2 | School of Electrical and Electronic Engineering |
| author_facet | School of Electrical and Electronic Engineering Karunasekera, Hasith Zhang, Handuo Wang, Han |
| author_sort | Karunasekera, Hasith |
| collection | NTU |
| description | In this paper we propose a tracking by detection method using a dissimilarity measure calculated based on the location and the appearance information of the object. These dissimilarity values are used in Hungarian Algorithm [1] in the data association step for track identity assignment. We make use of YOLO [2] deep learning based object detector in the detection step from camera image feed. Location measure is calculated using the predicted object location and bounding box, while the appearance measure is from the last feature layer from the detection network. Main focus in this work is to propose a tracking framework that can be used in real time automated vehicle guiding applications, by striking a balance between computational complexity and tracking accuracy. Therefore, we make use of the deep features available from detection framework rather than calculating a new appearance measure during the tracking step. The method proposed is very efficient and enables to achieve speeds up to 500+ frames per second (fps) in KITTI [3] tracking benchmark while achieving state-of-the-art results. |
| first_indexed | 2025-02-19T04:05:03Z |
| format | Conference Paper |
| id | ntu-10356/141797 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T04:05:03Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | ntu-10356/1417972020-06-11T00:44:54Z Real time multiple object tracking using deep features and localization information Karunasekera, Hasith Zhang, Handuo Wang, Han School of Electrical and Electronic Engineering 2019 IEEE 15th International Conference on Control and Automation (ICCA) Engineering::Electrical and electronic engineering MOT Computer Vision In this paper we propose a tracking by detection method using a dissimilarity measure calculated based on the location and the appearance information of the object. These dissimilarity values are used in Hungarian Algorithm [1] in the data association step for track identity assignment. We make use of YOLO [2] deep learning based object detector in the detection step from camera image feed. Location measure is calculated using the predicted object location and bounding box, while the appearance measure is from the last feature layer from the detection network. Main focus in this work is to propose a tracking framework that can be used in real time automated vehicle guiding applications, by striking a balance between computational complexity and tracking accuracy. Therefore, we make use of the deep features available from detection framework rather than calculating a new appearance measure during the tracking step. The method proposed is very efficient and enables to achieve speeds up to 500+ frames per second (fps) in KITTI [3] tracking benchmark while achieving state-of-the-art results. NRF (Natl Research Foundation, S’pore) Accepted version 2020-06-11T00:39:10Z 2020-06-11T00:39:10Z 2019 Conference Paper Karunasekera, H., Zhang, H., & Wang, H. (2019). Real time multiple object tracking using deep features and localization information. Proceedings of 2019 IEEE 15th International Conference on Control and Automation (ICCA), 332-337. doi:10.1109/ICCA.2019.8899498 9781728111650 https://hdl.handle.net/10356/141797 10.1109/ICCA.2019.8899498 2-s2.0-85075781648 332 337 en MRP1A © 2019 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/ICCA.2019.8899498. application/pdf |
| spellingShingle | Engineering::Electrical and electronic engineering MOT Computer Vision Karunasekera, Hasith Zhang, Handuo Wang, Han Real time multiple object tracking using deep features and localization information |
| title | Real time multiple object tracking using deep features and localization information |
| title_full | Real time multiple object tracking using deep features and localization information |
| title_fullStr | Real time multiple object tracking using deep features and localization information |
| title_full_unstemmed | Real time multiple object tracking using deep features and localization information |
| title_short | Real time multiple object tracking using deep features and localization information |
| title_sort | real time multiple object tracking using deep features and localization information |
| topic | Engineering::Electrical and electronic engineering MOT Computer Vision |
| url | https://hdl.handle.net/10356/141797 |
| work_keys_str_mv | AT karunasekerahasith realtimemultipleobjecttrackingusingdeepfeaturesandlocalizationinformation AT zhanghanduo realtimemultipleobjecttrackingusingdeepfeaturesandlocalizationinformation AT wanghan realtimemultipleobjecttrackingusingdeepfeaturesandlocalizationinformation |