Efficient Multi-Target Tracking using graphical models

Abstract

The objective of this thesis is to develop a new framework for Multi-Target Tracking (MTT) algorithms that are distinguished by the use of statistical machine learning techniques. MTT is a crucial problem for many important practical applications such as military surveillance. Despite being a well-studied research problem, MTT remains challenging, mostly because of the challenges of computational complexity faced by current algorithms. Taking a very di®erent approach from any existing MTT algorithms, we use the formalism of graphical models to model the MTT problem according to its probabilistic structure, and subsequently develop e±cient, approximate message passing algorithms to solve the MTT problem. Our modeling approach is able to take into account issues such as false alarms and missed detections. Although exact inference is intractable in graphs with a mix of both discrete and continuous random variables, such as the ones for MTT, our message passing algorithms utilize e±cient particle reduction techniques to make approximate inference tractable on these graphs. Experimental results show that our approach, while maintaining acceptable tracking quality, leads to linear running time complexity with respect to the duration of the tracking window. Moreover, our results demonstrate that, with the graphical model structure, our approach can easily handle special situations, such as out-of-sequence observations and track stitching.