Current Challenges with Modern Multi-Object Trackers - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2023

Current Challenges with Modern Multi-Object Trackers

Abstract

Multi-object tracking algorithms reach impressive performance on the benchmark datasets that they are trained and evaluated on, especially with their object detector parts tuned. When these algorithms are exposed to new videos though, the performance of the detection and tracking becomes poor, making them not usable. This paper tries to understand this behavior and discusses how we can move forward regarding these issues. Besides, we present the common errors made by the modern trackers, even when their trainable components are heavily tuned on the datasets as well as propose some directions and high-level ideas on how to proceed with those problems. We hope that it will lead to interesting discussions in the community towards solving such errors and further improving the performance of multi-object tracking.
Fichier principal
Vignette du fichier
Tomasz-ICCV-ACVR23.pdf (9.13 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04323242 , version 1 (06-12-2023)

Identifiers

  • HAL Id : hal-04323242 , version 1

Cite

Tomasz Stanczyk, Francois F Bremond. Current Challenges with Modern Multi-Object Trackers. ACVR 2023 - Eleventh International Workshop on Assistive Computer Vision and Robotics, Oct 2023, Paris, France. ⟨hal-04323242⟩
85 View
123 Download

Share

Gmail Mastodon Facebook X LinkedIn More