Performance evaluation of random set based pedestrian tracking algorithms
Paper in proceeding, 2013

The paper evaluates the error performance of three random finite set based multi-object trackers in the context of pedestrian video tracking. The evaluation is carried out using a publicly available video dataset of 4500 frames (town centre street) for which the ground truth is available. The input to all pedestrian tracking algorithms is an identical set of head and body detections, obtained using the Histogram of Oriented Gradients (HOG) detector. Head and body detections are unreliable in the sense that the probability of detection is low and false detections are non-uniformly distributed. The tracking error is measured using the recently proposed OSPA metric for tracks (OSPA-T), adopted as the only known mathematically rigorous metric for measuring the distance between two sets of tracks. The paper presents the correct proof of the triangle inequality for the OSPA-T. A comparative analysis is presented under various conditions.

Author

Branko Ristić

Defence Science and Technology Group

Jamie Sherrah

Defence Science and Technology Group

Angel Garcia

Chalmers, Signals and Systems, Signal Processing and Biomedical Engineering

2013 IEEE 8th International Conference on Intelligent Sensors, Sensor Networks and Information Processing: Sensing the Future, ISSNIP 2013

Vol. 1 300-305
978-146735500-1 (ISBN)

Subject Categories

Electrical Engineering, Electronic Engineering, Information Engineering

DOI

10.1109/ISSNIP.2013.6529806

ISBN

978-146735500-1

More information

Created

10/8/2017