Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker Fusion

This paper presents an approach to improving visual object tracking performance by dynamically fusing the results of two trackers, where the scheduling of trackers is determined by a support vector machine (SVM). By classifying the outputs of other trackers, our method learns their behaviors and exp...

Full description

Saved in:
Bibliographic Details
Main Authors: Vincenzo M. Scarrica, Antonino Staiano
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Technologies
Subjects:
Online Access:https://www.mdpi.com/2227-7080/12/12/239
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850060008404811776
author Vincenzo M. Scarrica
Antonino Staiano
author_facet Vincenzo M. Scarrica
Antonino Staiano
author_sort Vincenzo M. Scarrica
collection DOAJ
description This paper presents an approach to improving visual object tracking performance by dynamically fusing the results of two trackers, where the scheduling of trackers is determined by a support vector machine (SVM). By classifying the outputs of other trackers, our method learns their behaviors and exploits their complementarity to enhance tracking accuracy and robustness. Our approach consistently surpasses the performance of individual trackers within the ensemble. Despite being trained on only 4 sequences and tested on 144 sequences from the VOTS2023 benchmark, our approach achieves a Q metric of 0.65. Additionally, our fusion strategy demonstrates versatility across different datasets, achieving 73.7 MOTA on MOT17 public detections and 82.8 MOTA on MOT17 private detections. On the MOT20 dataset, it achieves 68.6 MOTA on public detections and 79.7 MOTA on private detections, setting new benchmarks in multi-object tracking. These results highlight the potential of using an ensemble of trackers with a learner-based scheduler to significantly improve tracking performance.
format Article
id doaj-art-ed2dead1ad6d4afb8e7e160e827f8b72
institution DOAJ
issn 2227-7080
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Technologies
spelling doaj-art-ed2dead1ad6d4afb8e7e160e827f8b722025-08-20T02:50:43ZengMDPI AGTechnologies2227-70802024-11-01121223910.3390/technologies12120239Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker FusionVincenzo M. Scarrica0Antonino Staiano1National PhD Program in AI—Agrifood and Environment, University of Naples Federico II, Corso Umberto I 40, 80138 Naples, ItalyDepartment of Science and Technology, University of Naples Parthenope, Centro Direrzione di Napoli, 80143 Naples, ItalyThis paper presents an approach to improving visual object tracking performance by dynamically fusing the results of two trackers, where the scheduling of trackers is determined by a support vector machine (SVM). By classifying the outputs of other trackers, our method learns their behaviors and exploits their complementarity to enhance tracking accuracy and robustness. Our approach consistently surpasses the performance of individual trackers within the ensemble. Despite being trained on only 4 sequences and tested on 144 sequences from the VOTS2023 benchmark, our approach achieves a Q metric of 0.65. Additionally, our fusion strategy demonstrates versatility across different datasets, achieving 73.7 MOTA on MOT17 public detections and 82.8 MOTA on MOT17 private detections. On the MOT20 dataset, it achieves 68.6 MOTA on public detections and 79.7 MOTA on private detections, setting new benchmarks in multi-object tracking. These results highlight the potential of using an ensemble of trackers with a learner-based scheduler to significantly improve tracking performance.https://www.mdpi.com/2227-7080/12/12/239multi-object trackingvisual object tracking segmentationhuman-crowd trackingsupport vector machine
spellingShingle Vincenzo M. Scarrica
Antonino Staiano
Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker Fusion
Technologies
multi-object tracking
visual object tracking segmentation
human-crowd tracking
support vector machine
title Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker Fusion
title_full Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker Fusion
title_fullStr Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker Fusion
title_full_unstemmed Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker Fusion
title_short Learning from Outputs: Improving Multi-Object Tracking Performance by Tracker Fusion
title_sort learning from outputs improving multi object tracking performance by tracker fusion
topic multi-object tracking
visual object tracking segmentation
human-crowd tracking
support vector machine
url https://www.mdpi.com/2227-7080/12/12/239
work_keys_str_mv AT vincenzomscarrica learningfromoutputsimprovingmultiobjecttrackingperformancebytrackerfusion
AT antoninostaiano learningfromoutputsimprovingmultiobjecttrackingperformancebytrackerfusion