Title :
A motion-enhanced hybrid Probability Hypothesis Density filter for real-time multi-human tracking in video surveillance scenarios
Author :
Eiselein, Volker ; Senst, Tobias ; Keller, Ivo ; Sikora, Thomas
Author_Institution :
Commun. Syst. Group, Tech. Univ. Berlin, Berlin, Germany
Abstract :
The Probability Hypothesis Density (PHD) filter is a multi-object Bayes filter which has been recently becoming popular in the tracking community especially for its linear complexity and its ability to filter out a high amount of clutter. However, its application to Computer Vision scenarios can be difficult as it requires high detection probabilities. Many human detectors suffer from a significant miss-match rate which causes problems for the PHD filter. This article presents an implementation of a Gaussian Mixture PHD (GM-PHD) filter which is enhanced by Optical Flow information in order to account for missed detections. We give a detailed mathematical discussion for the parameters of the proposed system and justify our results by extensive tests showing the performance in several contexts and on different datasets.
Keywords :
Bayes methods; Gaussian processes; computational complexity; filtering theory; motion estimation; object tracking; probability; real-time systems; video surveillance; GM-PHD filter; Gaussian Mixture PHD filter; PHD filter; computer vision; detection probabilities; human detectors; linear complexity; mathematical discussion; motion enhanced hybrid probability hypothesis density filter; multiobject Bayes filter; optical flow information; real-time multihuman tracking; tracking community; video surveillance scenarios; Computer vision; Detectors; Image motion analysis; Optical imaging; Optical sensors; Target tracking;
Conference_Titel :
Performance Evaluation of Tracking and Surveillance (PETS), 2013 IEEE International Workshop on
Conference_Location :
Clearwater, FL
Print_ISBN :
978-1-4673-5649-7
Electronic_ISBN :
2157-491X
DOI :
10.1109/PETS.2013.6523789
