FORT: Fisheye Online Realtime Tracking with an Improved Kalman Filter - IMT Mines Alès Access content directly
Conference Papers Year : 2023

FORT: Fisheye Online Realtime Tracking with an Improved Kalman Filter

Nathan Odic
  • Function : Author
Benoit Faure
  • Function : Author
  • PersonId : 1002549
Baptiste Magnier

Abstract

The goal of human tracking is to detect people in a scene and assign them a unique identifier that the tracker will follow across multiple frames. Our tracker, FORT, implements deep learning solutions such as, YOLOv7 for detection, ResNeXt- 50 for feature extraction and re-identification and an adapted Kalman filter for tracking. The goal is to present a real time tracking solution for the complex environment of top-view, fisheye images. The proposed solution is then compared with the BoTSORT and StrongSORT trackers on a custom fisheye Multiple Object Tracking (MOT) Challenge dataset
Fichier principal
Vignette du fichier
fort-fisheye-online-realtime-tracking.pdf (10.55 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04205295 , version 1 (10-06-2024)

Identifiers

Cite

Nathan Odic, Benoit Faure, Baptiste Magnier. FORT: Fisheye Online Realtime Tracking with an Improved Kalman Filter. MMSP 2023 - The IEEE International Workshop on MultiMedia Signal Processing, Sep 2023, Poitiers, France. ⟨10.1109/MMSP59012.2023.10337701⟩. ⟨hal-04205295⟩
60 View
5 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More