Multi-ROI Association and Tracking With Belief Functions: Application to Traffic Sign Recognition

Abstract

This paper presents an object tracking algorithm using belief functions applied to vision-based traffic sign recognition systems. This algorithm tracks detected sign candidates over time in order to reduce false positives due to data fusion formalization. In the first stage, regions of interest (ROIs) are detected and combined using the transferable belief model semantics. In the second stage, the local pignistic probability algorithm generates the associations maximizing the belief of each pairing between detected ROIs and ROIs tracked by multiple Kalman filters. Finally, the tracks are analyzed to detect false positives. Due to a feedback loop between the multi-ROI tracker and the ROI detector, the solution proposed reduces false positives by up to 45%, whereas computation time remains very low.

Coupled detection, association and tracking for Traffic Sign Recognition

“…Finally, even if a direct comparison of TSR performance rates is unrealistic [1], it can be shown that the MRT-based TSR provides interesting results with respect to state-of-theart solutions [19].…”

“…the reduction of false positive detections thanks to the joint association and tracking. An extended version of this paper providing additional experimental validations as well as a comparison with state-of-the-art systems is available in [19].…”

Evaluation of Probabilistic Transformations for Evidential Data Association

“…Several evidential data association approaches have been proposed [6,10,20,23] in the framework of belief functions. Rombaut [23] uses the Evidential theory to measure the confidence of the association between perceived and known obstacles.…”

“…This approach has been extended by Mercier et al [20] to track vehicles by using a global optimization to make assignment decisions. To reduce the complexity for real-time applications, a local optimization has been used [5,6]. For all these methods, the data fusion process begins by defining belief masses from sensor information and prior knowledge.…”

“…For data association applications, the widely used probabilistic transformation (i.e. approximation) is the pignistic transformation [5,6,17,20]. This transformation is based on a simple mapping process from belief to probability domain.…”