PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic Scenes
Multi-object tracking (MOT) is an important field in computer vision that provides a critical understanding of video analysis in various applications, such as vehicle tracking in intelligent transportation systems (ITS). Several deep learning-based approaches have been introduced to basic motion and...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9737499/ |
| _version_ | 1811271673401311232 |
|---|---|
| author | Ibrahim Soliman Mohamed Lim Kim Chuan |
| author_facet | Ibrahim Soliman Mohamed Lim Kim Chuan |
| author_sort | Ibrahim Soliman Mohamed |
| collection | DOAJ |
| description | Multi-object tracking (MOT) is an important field in computer vision that provides a critical understanding of video analysis in various applications, such as vehicle tracking in intelligent transportation systems (ITS). Several deep learning-based approaches have been introduced to basic motion and IoU trackers by extracting appearance features to assist in challenging situations such as lossy detection and occlusion. This study proposes a portable appearance extension (PAE) for single-stage object detection to jointly detect and extract appearance embeddings using a shared model. Furthermore, a novel training framework with a single image and without re-identification annotations is presented using an augmentation module, saving a tremendous amount of human labeling effort and increasing the real-world application adoption rate. Using UA-DETRAC dataset, RetinaNet-PAE and SSD-PAE achieve comparable results with current state-of-the-art models, where RetinaNet-PAE prioritizes detection and tracking performance with a 58.0% HOTA score and 4 FPS. In contrast, SSD-PAE prioritizes latency performance with a 47.3% HOTA score and 40 FPS. |
| first_indexed | 2024-04-12T22:24:08Z |
| format | Article |
| id | doaj.art-83c3d913975f462a8d46be461b6182ab |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-12T22:24:08Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-83c3d913975f462a8d46be461b6182ab2022-12-22T03:14:13ZengIEEEIEEE Access2169-35362022-01-0110372573726810.1109/ACCESS.2022.31604249737499PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic ScenesIbrahim Soliman Mohamed0https://orcid.org/0000-0003-4924-8851Lim Kim Chuan1https://orcid.org/0000-0002-0374-1993Machine Learning and Signal Processing, Centre for Telecommunication Research & Innovation, Faculty of Electronics and Computer Engineering, Universiti Teknikal Malaysia Melaka, Durian Tunggal, MalaysiaMachine Learning and Signal Processing, Centre for Telecommunication Research & Innovation, Faculty of Electronics and Computer Engineering, Universiti Teknikal Malaysia Melaka, Durian Tunggal, MalaysiaMulti-object tracking (MOT) is an important field in computer vision that provides a critical understanding of video analysis in various applications, such as vehicle tracking in intelligent transportation systems (ITS). Several deep learning-based approaches have been introduced to basic motion and IoU trackers by extracting appearance features to assist in challenging situations such as lossy detection and occlusion. This study proposes a portable appearance extension (PAE) for single-stage object detection to jointly detect and extract appearance embeddings using a shared model. Furthermore, a novel training framework with a single image and without re-identification annotations is presented using an augmentation module, saving a tremendous amount of human labeling effort and increasing the real-world application adoption rate. Using UA-DETRAC dataset, RetinaNet-PAE and SSD-PAE achieve comparable results with current state-of-the-art models, where RetinaNet-PAE prioritizes detection and tracking performance with a 58.0% HOTA score and 4 FPS. In contrast, SSD-PAE prioritizes latency performance with a 47.3% HOTA score and 40 FPS.https://ieeexplore.ieee.org/document/9737499/Computer visionobject detection and trackingappearance embeddingvehicle tracking |
| spellingShingle | Ibrahim Soliman Mohamed Lim Kim Chuan PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic Scenes IEEE Access Computer vision object detection and tracking appearance embedding vehicle tracking |
| title | PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic Scenes |
| title_full | PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic Scenes |
| title_fullStr | PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic Scenes |
| title_full_unstemmed | PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic Scenes |
| title_short | PAE: Portable Appearance Extension for Multiple Object Detection and Tracking in Traffic Scenes |
| title_sort | pae portable appearance extension for multiple object detection and tracking in traffic scenes |
| topic | Computer vision object detection and tracking appearance embedding vehicle tracking |
| url | https://ieeexplore.ieee.org/document/9737499/ |
| work_keys_str_mv | AT ibrahimsolimanmohamed paeportableappearanceextensionformultipleobjectdetectionandtrackingintrafficscenes AT limkimchuan paeportableappearanceextensionformultipleobjectdetectionandtrackingintrafficscenes |