Blob trackers have become increasingly powerful in recent years largely due to the adoption of statistical appearance models which allow effective background subtraction and robust tracking of deforming foreground objects. It has been standard, however, to treat background and foreground modelling as separate processes --- background subtraction is followed by blob detection and tracking --- which prevents a principled computation of image likelihoods. This paper presents two theoretical advances which address this limitation and lead to a robust multiple-person tracking system suitable for single-camera real-time surveillance applications. The first innovation is a multi-blob likelihood function which assigns directly comparable likelihoods to hypotheses containing different numbers of objects. This likelihood function has a rigorous mathematical basis: it is adapted from the theory of Bayesian correlation, but uses the assumption of a static camera to create a more specific background model while retaining a unified approach to background and foreground modelling. Second we introduce a Bayesian filter for tracking multiple objects when the number of objects present is unknown and varies over time. We show how a particle filter can be used to perform joint inference on both the number of objects present and their configurations. Finally we demonstrate that our system runs comfortably in real time on a modest workstation when the number of blobs in the scene is small.

Abstract not found

This thesis establishes a stochastic framework for tracking curves in visual clutter, using a Bayesian random-sampling algorithm. The approach is rooted in ideas from statistics, control theory and computer vision. The problem is to track outlines and features of foreground objects, modelled as curves, as they move in substantial clutter, and to do it at, or close to, video frame-rate. The algorithm, named Condensation, for Conditional density propagation, has recently been derived independently by several researchers, and is generating signi cant interest in the statistics and signal processing communities. This thesis contributes to the literature on Condensation-like lters by presenting some novel applications of and extensions to the basic algorithm, and contributes to the visual motion estimation literature by demonstrating high tracking performance in cluttered environments. Despite its power the Condensation algorithm has a remarkably simple form and this allows the use of non-linear motion models which combine characteristics of discrete Hidden Markov Models with the continuous Auto-Regressive Process motion models traditionally used in Kalman lters. These mixed discrete-continuous models have promising applications to the emerging eld of perception of action. This thesis also implements two algorithms to smooth the output of the Condensation lter which improves the accuracy of motion estimation in a batch-mode procedure after tracking is complete.

Abstract. A method for matching sequences from two perspective views of a moving person silhouette is presented. Regular (approximate uniform thickness) parts are detected on an image and a skeleton is generated. A 3D regular region graph is defined to gather possible poses based on the two 2D-regular regions, one for each view, at a given frame. The matching process of 3D graphs with a model graph results in interpretations of the human motion in the scene. The objective of this system is to reconstruct human motion parameters and use the analytical information for synthesis. Experimental results and error analysis are explained when the system is used to drive an avatar. 1

Abstract — Today, more than ever, monitoring and surveillance systems play an important role in many aspects of our lives. Technology plays a vital role in our efforts to create, store and analyze vast amounts of data for both security and commercial purposes. In this paper, we propose an application which combines research performed in computer networks, multimedia databases and computer vision. We consider the problem where a number of networks are interconnected. Each of the individual nodes (networks) are collecting, processing and storing data from several sensors (cameras). Specifically, we emphasize on how the data (images) are processed by the individual nodes and how the information is transmitted, so that queries involving multiple nodes can be answered. During this process, we also identify several challenges related to sharing voluminous content provided by visual surveillance devices.