Pedestrian multiple-object tracking based on FairMOT and circle loss
Abstract Multi-object Tracking is an important issue that has been widely investigated in computer vision. However, in practical applications, moving targets are often occluded due to complex changes in the background, which leads to frequent pedestrian ID switches in multi-object tracking. To solve...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Nature Portfolio
2023-03-01
|
Series: | Scientific Reports |
Online Access: | https://doi.org/10.1038/s41598-023-31806-2 |
_version_ | 1797864650094673920 |
---|---|
author | Jin Che Yuting He Jinman Wu |
author_facet | Jin Che Yuting He Jinman Wu |
author_sort | Jin Che |
collection | DOAJ |
description | Abstract Multi-object Tracking is an important issue that has been widely investigated in computer vision. However, in practical applications, moving targets are often occluded due to complex changes in the background, which leads to frequent pedestrian ID switches in multi-object tracking. To solve the problem, we present a multi-object tracking algorithm based on FairMOT and Circle Loss. In this paper, HRNet is adopted as the baseline. Then, Polarized Self-Attention is added to HRNet-w32 to obtain weights of helpful information based on its modeling advantages. Moreover, the re-identification branch is optimized, and the Circle Loss is selected as the loss function to acquire more discriminative pedestrian features and to distinguish different pedestrians. The method proposed is assessed on the public MOT17 datasets. The experimental results show that the MOTA score achieves 69.5%, IDF1 reaches 70.0%, and the number of ID switches (IDs) decreases 636 times compared to the TraDes algorithm. |
first_indexed | 2024-04-09T22:56:26Z |
format | Article |
id | doaj.art-ad23e3fa9132430e882f09da06ebec8c |
institution | Directory Open Access Journal |
issn | 2045-2322 |
language | English |
last_indexed | 2024-04-09T22:56:26Z |
publishDate | 2023-03-01 |
publisher | Nature Portfolio |
record_format | Article |
series | Scientific Reports |
spelling | doaj.art-ad23e3fa9132430e882f09da06ebec8c2023-03-22T11:16:26ZengNature PortfolioScientific Reports2045-23222023-03-0113111210.1038/s41598-023-31806-2Pedestrian multiple-object tracking based on FairMOT and circle lossJin Che0Yuting He1Jinman Wu2School of Physics and Electronic-Electrical Engineering, Ningxia UniversitySchool of Physics and Electronic-Electrical Engineering, Ningxia UniversitySchool of Physics and Electronic-Electrical Engineering, Ningxia UniversityAbstract Multi-object Tracking is an important issue that has been widely investigated in computer vision. However, in practical applications, moving targets are often occluded due to complex changes in the background, which leads to frequent pedestrian ID switches in multi-object tracking. To solve the problem, we present a multi-object tracking algorithm based on FairMOT and Circle Loss. In this paper, HRNet is adopted as the baseline. Then, Polarized Self-Attention is added to HRNet-w32 to obtain weights of helpful information based on its modeling advantages. Moreover, the re-identification branch is optimized, and the Circle Loss is selected as the loss function to acquire more discriminative pedestrian features and to distinguish different pedestrians. The method proposed is assessed on the public MOT17 datasets. The experimental results show that the MOTA score achieves 69.5%, IDF1 reaches 70.0%, and the number of ID switches (IDs) decreases 636 times compared to the TraDes algorithm.https://doi.org/10.1038/s41598-023-31806-2 |
spellingShingle | Jin Che Yuting He Jinman Wu Pedestrian multiple-object tracking based on FairMOT and circle loss Scientific Reports |
title | Pedestrian multiple-object tracking based on FairMOT and circle loss |
title_full | Pedestrian multiple-object tracking based on FairMOT and circle loss |
title_fullStr | Pedestrian multiple-object tracking based on FairMOT and circle loss |
title_full_unstemmed | Pedestrian multiple-object tracking based on FairMOT and circle loss |
title_short | Pedestrian multiple-object tracking based on FairMOT and circle loss |
title_sort | pedestrian multiple object tracking based on fairmot and circle loss |
url | https://doi.org/10.1038/s41598-023-31806-2 |
work_keys_str_mv | AT jinche pedestrianmultipleobjecttrackingbasedonfairmotandcircleloss AT yutinghe pedestrianmultipleobjecttrackingbasedonfairmotandcircleloss AT jinmanwu pedestrianmultipleobjecttrackingbasedonfairmotandcircleloss |