Single-Task Joint Learning Model for an Online Multi-Object Tracking Framework

Single-Task Joint Learning Model for an Online Multi-Object Tracking Framework

Multi-object tracking faces critical challenges, including occlusions, ID switches, and erroneous detection boxes, which significantly hinder tracking accuracy in complex environments. To address these issues, this study proposes a single-task joint learning (STJL) model integrated into an online mu...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuan-Kai Wang, Tung-Ming Pan, Chi-En Hu
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Applied Sciences
Subjects:
multi-object tracking
single-task joint learning
cross-dataset training
feature extraction
tracker initialization
cosine distance
Online Access:https://www.mdpi.com/2076-3417/14/22/10540
Tags: Add Tag
No Tags, Be the first to tag this record!

Internet

https://www.mdpi.com/2076-3417/14/22/10540

Similar Items