Christopher S. Madden

Tracking people across disjoint camera views by an illumination-tolerant appearance representation

Tracking single individuals as they move across disjoint camera views is a challenging task since their appearance may vary significantly between views. Major changes in appearance are due to different and varying illumination conditions and the deformable geometry of people. These effects are hard to estimate and take into account in real-life applications. Thus, in this paper we propose an illumination-tolerant appearance representation, which is capable of coping with the typical illumination changes occurring in surveillance scenarios. The appearance representation is based on an online k-means colour clustering algorithm, a data-adaptive intensity transformation and the incremental use of frames. A similarity measurement is also introduced to compare the appearance representations of any two arbitrary individuals. Post-matching integration of the matching decision along the individuals‘ tracks is performed in order to improve reliability and robustness of matching. Once matching is provided for any two views of a single individual, its tracking across disjoint cameras derives straightforwardly. Experimental results presented in this paper from a real surveillance camera network show the effectiveness of the proposed method.

Measuring Latency for Video Surveillance Systems

The increased flexibility and other benefits offered by IP network cameras makes them a common choice for installation in new and expanded surveillance networks. One commonly quoted limitation of IP cameras is their high latency when compared to their analogue counterparts. This causes some reluctance to install or upgrade to digital cameras, and is slowing the adoption of live, intelligent analysis techniques in video surveillance systems. This paper presents methods for measurement of the latency in systems based upon digital IP or analogue cameras. These methods are camera-agnostic and require no specialised hardware. We use these methods to compare a variety of camera models. The results demonstrate that whilst analogue cameras do have a lower latency, most IP cameras are within acceptable tolerances. The source of the latency within an IP camera is also analysed, with prospects for improvement identified.

A framework for track matching across disjoint cameras using robust shape and appearance features

This paper presents a framework based on robust shape and appearance features for matching the various tracks generated by a single individual moving within a surveillance system. Each track is first automatically analysed in order to detect and remove the frames affected by large segmentation errors and drastic changes in illumination. The objects features computed over the remaining frames prove more robust and capable of supporting correct matching of tracks even in the case of significantly disjointed camera views. The shape and appearance features used include a height estimate as well as illumination-tolerant colour representation of the individuals global colours and the colours of the upper and lower portions of clothing. The results of a test from a real surveillance system show that the combination of these four features can provide a probability of matching as high as 91 percent with 5 percent probability of false alarms under views which have significantly differing illumination levels and suffer from significant segmentation errors in as many as 1 in 4 frames.

Automatic camera placement for large scale surveillance networks

Automatic placement of surveillance cameras in arbitrary buildings is a challenging task, and also one that is essential for efficient deployment of large scale surveillance networks. Existing approaches for automatic camera placement are either limited to a small number of cameras, or constrained in terms of the building layouts to which they can be applied. This paper describes a new method for determining the best placement for large numbers of cameras within arbitrary building layouts. The method takes as input a 3D model of the building, and uses a genetic algorithm to find a placement that optimises coverage and (if desired) overlap between cameras. Results are reported for an implementation of the method, including its application to a wide variety of complex buildings, both real and synthetic.

Tracking hand-off in large surveillance networks

This paper investigates the use of pairwise camera overlap estimates for supporting target tracking across large networks of surveillance cameras. We compare the use of camera overlap topology information to a method based on matching target appearance histograms, and also evaluate the effect of combining both methods. Tracking accuracy results are reported in terms of precision and recall for a 24 camera network. Camera overlap information is shown to deliver significant advantages for tracking when compared with simply matching target appearance histograms, due to its robustness to low quality imagery. We show empirically that this is the case even for automatically derived overlap estimates containing errors.

Mitigating the Effects of Variable Illumination for Tracking across Disjoint Camera Views

Tracking people by their appearance across disjoint camera views is challenging since appearance may vary significantly across such views. This problem has been tackled in the past by computing intensity transfer functions between each camera pair during an initial training stage. However, in real-life situations, intensity transfer functions depend not only on the camera pair, but also on the actual illumination at pixel-wise resolution and may prove impractical to estimate to a satisfactory extent. For this reason, in this paper we propose an appearance representation for people tracking capable of coping with the typical illumination changes occurring in a surveillance scenario. Our appearance representation is based on an online K-means color clustering algorithm, a fixed, data-dependent intensity transformation, and the incremental use of frames. Moreover, a similarity measurement is proposed to match the appearance representations of any two given moving objects along sequences of frames. Experimental results presented in this paper show that the proposed methods provides a viable while effective approach for tracking people across disjoint camera views in typical surveillance scenarios.

Comparison of techniques for mitigating the effects of illumination variations on the appearance of human targets

Several techniques have been proposed to date to build colour invariants between camera views with varying illumination conditions. In this paper, we propose to improve colour invariance by using data-dependent techniques. To this aim, we compare the effectiveness of histogram stretching, illumination filtration, full histogram equalisation and controlled histogram equalisation in a video surveillance domain. All such techniques have limited computational requirements and are therefore suitable for real time implementation. Controlled histogram equalisation is a modified histogram equalisation operating under the influence of a control parameter [1]. Our empirical comparison looks at the ability of these techniques to make the global colour appearance of single human targets more matchable under illumination changes, whilst still discriminating between different people. Tests are conducted on the appearance of individuals from two camera views with greatly differing illumination conditions and invariance is evaluated through a similarity measure based upon colour histograms. In general, our results indicate that these techniques improve colour invariance; amongst them, full and controlled equalisation consistently showed the best performance.

A framework for determining overlap in large scale networks

This paper presents a novel framework designed for calculating the topology of overlapping cameras in large surveillance systems. Such a framework is a key enabler for efficient network-wide surveillance, e.g. inter-camera tracking, especially in large surveillance networks. The framework presented can be adapted to utilise numerous contradiction and correlation approaches to identify overlapping portions of camera views using activity within the system. It can also utilise a various arbitrary occupancy cells which can be used to adjust both the memory requirements and accuracy of the topology generated. The framework is evaluated for its memory usage, processing speed and the accuracy of its overlap topology on a 26 camera dataset using various approaches. A further examination of memory requirements and processing speed on a larger 200 camera network is also presented. The results demonstrate that the framework significantly reduces memory requirements and improves execution speed whilst producing useful topologies from a large surveillance system at real-time speeds.

Surprisal-aware scheduling of PTZ cameras

An approach is presented for scheduling PTZ cameras on guard tours with two or more fields of view. In contrast to the target tracking of previous work, this approach seeks to optimise the coverage of the area under surveillance. Specifically, the aim is to minimise the surprisal (self-information) of events in unobserved fields of view. An entropy driven scheduler based on Kullback-Leibler divergence (information gain) is presented, and compared with three naive schedulers (random, round robin and constant selection of one field of view). Experiments investigate its performance on networks of ten cameras. These are evaluated over factors including four different scheduling approaches, different numbers of fields of view, and different inactive times whilst switching views. They demonstrate the efficacy of the entropy driven scheduler as it outperforms the naive schedulers by a significant margin by favouring certain fields of view that are more likely to reveal events with high surprisal value. The scheduler is target agnostic, as it operates on low level properties of the video signal, specifically, occupancy as determined by background subtraction. This permits an efficient implementation that is independent of the number of targets in the area under surveillance. As each camera is scheduled independently, the approach is scalable via distributed implementation, including on smart cameras.

Contradiction and Correlation for Camera Overlap Estimation

An accurate estimate of camera overlap is a key enabler for efficient network-wide surveillance processing (e.g. inter-camera tracking), especially in large-scale surveillance networks. Techniques based on contradictions in pair-wise occupancy data, such as the exclusion approach, have advantages in robustness and efficiency that make them particularly well suited for large surveillance networks. Correlation techniques share some of these advantages,but have a better understood statistical basis. This paper evaluates a set of contradiction and correlation techniques, using a novel metric, search space precision-recall. This metric reflects the activity-based overlap estimation required for camera handover, such as would be used in inter-camera tracking.Results are reported for a range of networks, including a 24-camera network setup in an office space, where the exclusion estimator showed the best performance.