Sze Ling Tang

iSurveillance: Intelligent framework for multiple events detection in surveillance videos

Abstract Research in the video surveillance is gaining more popularity due to its widespread applications as well as social impact. In this paper, we present an intelligent framework for detection of multiple events in surveillance videos. Based on the principle of compositionality, we modularize the surveillance problems into a set of variables comprising regions-of-interest, classes (i.e. human, vehicle), attributes (i.e. speed, locality) and a set of notions (i.e. rules) associated to each of the attributes to construct a knowledge-based understanding of the environment. The final output from the reasoning process, which combines the definition domains of the various variables, allows a broader and integrated understanding of complex pattern of activities in the scene. This is in contrast to the state-of-the-art solutions that are only able to perform only a singular task, at a time. Experimental results on both the public and real-time datasets have demonstrated the effectiveness and robustness of the proposed framework in detecting multiple events in surveillance videos.

Hybrid blob and particle filter tracking approach for robust object tracking

Abstract Analysing and characterising human behaviour is now receiving much attention from the visual surveillance research community. Generally, human behaviour recognition requires human to be detected and tracked so that the trajectory patterns of the human can be captured and analysed for further interpretation. Therefore, it is crucial for tracking algorithms to be fast and robust to partial and short-life occlusion. In addition, the detection of object-of-interest to be tracked should be automated, without the need for manual intervention. This paper thus proposes a tracking system targeted for real time surveillance applications that integrate blob and simplified particle filter tracking approaches so as to exploit the advantages of both approaches while minimizing their respective disadvantages. The blob approach acts as the main tracking and will invoke the simplified particle filter tracking in the event of blob merging or occlusion. In this paper, the proposed tracking method is tested using PETS 2009 sequences to illustrate the capability of solving occlusion and obstruction in the scene. The results show that the proposed system successfully tracks objects during and after occlusion with other objects or after obstructed by the background.

Orthogonal planar search (OPS) for coronary artery centerline extraction

In this paper, we investigate orthogonal planar search (OPS) for coronary artery centerline extraction to assist in coronary artery diseases diagnosis. The search mechanism exploits a data-driven algorithm to extract the centerline. Firstly, the best representation of vessel cross section on orthogonal planar is determined. Then, the center of gravity from the crosssection is computed as centerline point iteratively. Branching detection and termination are invoked in this proposed method. We demonstrate the results quantitatively and qualitatively. In addition, we benchmark OPS with three state-of-the-art methods and illustrate the comparison results in radar chart (also known as spider chart). Finally, we discuss limitations of OPS and future works.