Event Composition with Imperfect Information for Bus Surveillance

2007

Video surveillance is the process of monitoring the behavior of people and objects within public places, e.g. airports and traffic intersections, by means of visual aids (cameras) usually for safety and security purposes. As the amount of video data gathered daily by surveillance cameras increases, the need for automatic systems to detect and recognize suspicious activities performed by people and objects is also increasing. The first part of the thesis describes a framework for modeling and recognition of events from surveillance video. Our framework is based on deterministic inference using Petri nets. Events can be composed by combining primitive events and previously defined events by spatial, temporal and logical relations. We provide a graphical user interface (GUI) to formulate such event models. Our approach automatically maps each of these models into a set of Petri net filters that represent the components of the event. Last but definitely not least, I would like to thank my two lovely daughters; Farah and Sarah for their unconditional love and patience during completing this thesis. iv TABLE OF CONTENTS List of Figures ix above and to the left of (x; y), inclusive. (b) The sum of the pixels within rectangle D is computed as: ii(4) + ii(1) − ii(2) − ii(3). .. .. .. . 4.4 Examples from Backpack Dataset Showing Different Ways of Holding the Backpack .

Event representation models provide a framework in which we can reason about events so as to interpret the collective behaviour of objects over time and space domains. Many are context-specific and lack flexibility when faced with unstructured video. In the past many efforts have been made to define a comprehensive event description framework (EDF), which can provide a framework to develop ontologies for semantic annotation of video events. However, it is observed that there are some areas of event modelling that were not fully explored. Hence, we extended and modified the EDF and proposed the extended version of it (EDF E). Following are some of the major extensions we have proposed in EDF E. 1) EDF E extends the entity representation model of EDF by introducing three new entity classes: that of text entity, virtual entity and internal entity. II) EDF E introduces a new set of predicates for describing more complex event scenarios and facilitating the event detection process. It also introduces granularity as a feature of temporal predicates to capture the temporal association between sub-events. III) It introduces the event evidence feature to capture the full evidence for the detected events. IV) The data structure of EDF is extended and modified to capture the properties of EDF E and to store the results of the event detection process. V) We model complex events from real world surveillance videos using the proposed EDF E .

Image and Vision Computing, 2009

In this paper, we propose a technique for detecting possible events in outdoor areas monitored by a video surveillance system. In particular, here we focus on the time spent by an object to carry out simple events. To have a statistical representation of the times commonly required to perform certain activities, mixtures of Gaussians are maintained for each event type. Such statistics are then exploited both for the analysis of the simple activities and for discovering anomalous situations (i.e. complex events). In these cases, the system requires the attention of the human operator. A novel way of presenting results to the operator is also discussed. Experiments have been performed on a multi-camera system for parking lot security.

Lecture Notes in Computer Science, 2008

We address the problem of visual event recognition in surveillance where noise and missing observations are serious problems. Common sense domain knowledge is exploited to overcome them. The knowledge is represented as first-order logic production rules with associated weights to indicate their confidence. These rules are used in combination with a relaxed deduction algorithm to construct a network of grounded atoms, the Markov Logic Network. The network is used to perform probabilistic inference for input queries about events of interest. The system's performance is demonstrated on a number of videos from a parking lot domain that contains complex interactions of people and vehicles.

Lecture Notes in Computer Science, 2011

In this paper we fully develop a fall detection application that focuses on complex event detection. We use a decoupled approach, whereby the definition of events and of their complexity is fully detached from low and intermediate image processing level. We focus on context independence and flexibility to allow the reuse of existing approaches on recognition task. We build on existing proposals based on domain knowledge representation through ontologies. We encode knowledge at the rule level, thus providing a more flexible way to handle complexity of events involving more actors and rich time relationships. We obtained positive results from an experimental dataset of 22 recordings, including simple and complex fall events.

IET Computer Vision, 2008

The authors propose a high-level scenario recognition algorithm for video sequence interpretation. The recognition of scenarios is based on a Bayesian networks approach. The model of a scenario contains two main layers. The first one allows events from the observed visual features to be highlighted and the second layer is focused on the temporal reasoning stage. The temporal layer uses specific nodes permitting an event-based approach. These nodes focus on the lifetime of events highlighted from the results of the first layer. The temporal layer then estimates the qualitative and quantitative relations between the different temporal events helpful for the recognition task. The global recognition algorithm is illustrated over real indoor image sequences of an abandoned baggage scenario.

"International Workshop on Distributed Event-based Systems (DEBS 2004)" W18L Workshop - 26th International Conference on Software Engineering, 2004

In this paper we are concerned with event-based situation analysis. Application areas include the understanding and awareness of complex unfolding scenarios such as homeland security threats and future battlespace engagements. The paper (i) discusses the differences between the environments/requirements for event-based management and situation management, (ii) presents an argument for an integration of an event correlation system and a situation awareness system, and (iii) proposes an integrated architecture that combines rule-based spatio-temporal event correlation and casebased reasoning for understanding and managing situations. In addition, the paper discusses the hard problem of the mutual influence between event management and situation awareness.

2009 IEEE International Conference on Data Mining Workshops, 2009

Advances in RFID based sensor technologies has been used in applications which requires the tracking of assets, products and individuals. The recording of such movements is captured in a trajectory database and can be analyzed for the monitoring of abnormal events. In this paper, we describe a system called InViTA for analyzing abnormal events from spatio-temporal trajectories captured during an of ce evacuation after an explosion. InViTA utilizes a trajectory representation scheme and extract the features to derive a set of rules that label each person's trajectory as belonging to a suspect, witness, or victim, etc. We run the system on the of ce evacuation data provided in VAST 2008 challenge and obtain comparable results with that obtained from visualization and human analysis. The system includes a user-friendly graphical interface for parameter tuning and intuitive result analysis.

Expert Systems with Applications

In the UK alone there are currently over 4.2 million operational CCTV cameras, that is virtually one camera for every 14th person, and this figure is increasing at a fast rate throughout the world (especially after the tragic events of 9/11 and 7/7) (Norris, McCahill, & Wood, 2004). Security concerns are not the only factor driving the rapid growth of CCTV cameras. Another important reason is the access of hidden knowledge extracted from CCTV footage to be used for effective business decision making, such as store designing, customer services, product marketing, reducing store shrinkage, etc.Events occurring in observed scenes are one of the most important semantic entities that can be extracted from videos (Anwar & Naftel, 2008). Most of the work presented in the past is based upon finding frequent event patterns or deals with discovering already known abnormal events. In contrast, in this paper we present a framework to discover unknown anomalous events associated with a frequent ...

Peerj Computer Science, 2022

Digital vision technologies emerged exponentially in all living areas to watch, play, control, or track events. Security checkpoints have benefited also from those technologies by integrating dedicated cameras in studied locations. The aim is to manage the vehicles accessing the inspection security point and fetching for any suspected ones. However, the gathered data volume continuously increases each day, making their analysis very hard and time-consuming. This paper uses semantic-based techniques to model the data flow between the cameras, checkpoints, and administrators. It uses ontologies to deal with the increased data size and its automatic analysis. It considers forensics requirements throughout the creation of the ontology modules to ensure the records' admissibility for any possible investigation purposes. Ontology-based data modeling will help in the automatic events search and correlation to track suspicious vehicles efficiently.