@InProceedings{2015_craciun184,
	author = "Paula Craciun and Mathias Ortner and Josiane Zerubia",
	title = "Integrating RJMCMC and Kalman filters for multiple object tracking",
	booktitle = "25° Colloque sur le traitement du signal et des images",
	year = "2015",
	publisher = "GRETSI - Groupe de Recherche en Traitement du Signal et des Images",
	number = "001-0134",
	pages = "p. 537-540",
	month = "Sep # 8--11",
	address = "Lyon",
	doi = "",
	pdf = "2015_craciun184.pdf",
	abstract = "In this paper, we propose to integrate the Kalman filter with the reversible jump Markov Chain Monte Carlo (RJMCMC) sampler to improve the optimization procedure in the case of multiple object tracking. We propose the use of a dedicated perturbation kernel that uses the Kalman filter to generate multiple objects in a single iteration. We demonstrate that this kernel reduces considerably the mixing time of the Markov chain, as compared to the standard RJMCMC sampler. We show results on two synthetic biological sequences and two simulated remotely sensed data sets of the city of Toulon, France..pdf",
}