Enregistré dans:
Détails bibliographiques
Auteurs principaux: Haag, Stefan, Duraisamy, Bharanidhar, Govaers, Felix, Koch, Wolfgang, Fritzsche, Martin, Dickmann, Juergen
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2508.09585
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909735732641792
author Haag, Stefan
Duraisamy, Bharanidhar
Govaers, Felix
Koch, Wolfgang
Fritzsche, Martin
Dickmann, Juergen
author_facet Haag, Stefan
Duraisamy, Bharanidhar
Govaers, Felix
Koch, Wolfgang
Fritzsche, Martin
Dickmann, Juergen
contents This paper introduces BAAS, a new Extended Object Tracking (EOT) and fusion-based label annotation framework for radar detections in autonomous driving. Our framework utilizes Bayesian-based tracking, smoothing and eventually fusion methods to provide veritable and precise object trajectories along with shape estimation to provide annotation labels on the detection level under various supervision levels. Simultaneously, the framework provides evaluation of tracking performance and label annotation. If manually labeled data is available, each processing module can be analyzed independently or combined with other modules to enable closed-loop continuous improvements. The framework performance is evaluated in a challenging urban real-world scenario in terms of tracking performance and the label annotation errors. We demonstrate the functionality of the proposed approach for varying dynamic objects and class types
format Preprint
id arxiv_https___arxiv_org_abs_2508_09585
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Offline Auto Labeling: BAAS
Haag, Stefan
Duraisamy, Bharanidhar
Govaers, Felix
Koch, Wolfgang
Fritzsche, Martin
Dickmann, Juergen
Computer Vision and Pattern Recognition
Systems and Control
This paper introduces BAAS, a new Extended Object Tracking (EOT) and fusion-based label annotation framework for radar detections in autonomous driving. Our framework utilizes Bayesian-based tracking, smoothing and eventually fusion methods to provide veritable and precise object trajectories along with shape estimation to provide annotation labels on the detection level under various supervision levels. Simultaneously, the framework provides evaluation of tracking performance and label annotation. If manually labeled data is available, each processing module can be analyzed independently or combined with other modules to enable closed-loop continuous improvements. The framework performance is evaluated in a challenging urban real-world scenario in terms of tracking performance and the label annotation errors. We demonstrate the functionality of the proposed approach for varying dynamic objects and class types
title Offline Auto Labeling: BAAS
topic Computer Vision and Pattern Recognition
Systems and Control
url https://arxiv.org/abs/2508.09585