The School of Informatics has been awarded funding from the EPSRC for a project entitled "BEHAVE: Computed-assisted prescreening of video streams for unusual activities". The principal investigator on the project is Dr. Robert Fisher. The project funding is until September 30, 2007.
The research proposed here will take place in the Machine Vision Unit of the Institute of Perception, Action and Behaviour (http://www.ipab.informatics.ed.ac.uk). The Institute investigates how to link, in theory and in practice, computational action, perception, representation, transformation and generation processes to external worlds. Research activities include computer vision, mobile and assembly robotics and visualization. The Machine Vision Unit has a long history of research with three-dimensional data, including 3D sensing, surface description, object recognition and automatic model acquisition. More details can be found at http://www.ipab.informatics.ed.ac.uk/mvu/. In the area of this project, previous research in the Machine Vision Unit has demonstrated successful foveal feature extraction and appearance-based 2D object recognition.
The Institute of Perception, Action and Behaviour is a research institute in the School of Informatics at Edinburgh University. The University is internationally known for its School of Informatics, which is now the largest category 5 academic research and teaching department of Informatics in the United Kingdom. At present there are about 65 academic staff, 55 research staff and about 250 postgraduate MSc and PhD students. Edinburgh was the only University in the UK awarded the top 5*A rating in Computer Science in the 2001 Research Assessment Exercise. Edinburgh is also the UK's biggest research group in this area. Research activities cover almost every aspect of Informatics. More details about the School and its research can be found here.
The research proposed here is funded by the UK's EPSRC (Engineering and Physical Science Research Council) Think Crime initiative.
The project will investigate two novel computer-based image analysis processes to prescreen video sequences for abnormal or crime-oriented behaviour. The ultimate goal is to filter out image sequences where uninteresting normal activity is occurring, as well as the much easier sequences where nothing is occurring. The first process is for detecting, understanding and discriminating between similar types of interactions, such as two people fighting versus meeting and greeting. We propose to investigate and possibly extend the dynamic Hidden Markov Model technique as applied to tracked individuals to solve this problem. The second process is for analysing crowd scenes, where tracking of individuals is only possible over short time periods, and where the overall flow of the crowd is more salient. The goal is to discriminate between normal behaviour, such as people normally exiting from a football match, and abnormal behaviour, such as when people have to divert around an obstacle (fallen person, fight, etc). We propose to adapt global probabilistic models to flow data obtainable from short-time image tracking.
The objectives are:
With the recent installation of video surveillance systems in many city centres and other urban areas, there is a massive increase in the ability to collect data. Traditionally, this data was processed by human observers, recently replaced by recording equipment and post-processed only after undesirable events have occurred. What is more desirable is to be able to automatically detect potentially significant events as they happen. This full capability is beyond current computer technology and still requires human observers. Unfortunately, there are not enough operators and boredom also sets in quickly.
The research proposed here is aimed at collaborative working between human observers and computer-assisted prescreening: the computer would make initial assessments of video streams to select interesting sequences, which are then switched through to human operators for their more subtle assessment. This also allows a single human to manage more cameras simultaneously, as only the significant data will be relayed. Moreover, the portions of the video stream where the interesting events are occurring can be highlighted.
From a pragmatic viewpoint, the results could extend the capabilities of surveillance ``blank-screen'' technology, wherein operators are only presented with information when activity is occurring in the scene. The approaches presented here could also eliminate routine normal activity, such as two people walking together, so as to allow human operators to focus on unusual behaviours.
At the moment, there are a few good results on modelling discrete interactions using pure probabilistic and mixed probability and symbolic representations. What is less well understood is how to model subtle distinctions between slightly different types of interactions, such as greeting behaviour versus fighting. In the area of recognising behaviour in contexts with many actors, such as crowd behaviour at sporting events, we have not found any good prior work. Thus, we claim that there are two good open research questions here, both of which have direct relevance to the crime detection and prevention programme.
If successful, the research proposed here will:
The project will investigate two sub-problems based on interacting groups of humans.
The goal of this subproject is to extend existing methods of behaviour recognition to be able to distinguish between subtly different interactions between a small group of individuals, in particular between greeting and fighting, or preparing for fighting.
When the number of interacting people increases beyond a threshold level, the performance of tracking individuals will decrease, and thus symbolic interpretation of the behaviour becomes impossible. Therefore, we propose investigating a novel flow-based approach. In this approach, short-term correlation-based tracking can produce flow patterns in the image data. From these patterns, statistical classification techniques can probably be developed that distinguish between normal and abnormal flow patterns. For example, fans leaving a football ground normally have standard movement patterns, leading to standard flow patterns. If the flow is disrupted, e.g. by a fight, then crowd density may make it impossible to track individuals or identify the fighters. However, the disruption to the flow because of obstacles and other people attempting to avoid the fighters may become detectable.
The project will use several PCs attached to a file server. Image data will be by attached colour web cameras. Altogether, there are 10 workstations available for use by the vision research group (consisting of about 10 members, including contract research staff, PhD and MSc students). A 25 node Beowulf parallel system is also part of the group equipment. Wherever possible we will use either MATLAB, C, C++ or JAVA within a LINUX/UNIX environment (mainly for speed). There is some existing software related to this project.
There are several other people researching similar topics, namely in the CAVIAR: Context Aware Vision using Image-based Active Recognition project. The main objective of CAVIAR is to address the scientific question: Can rich local image descriptions from foveal and other image sensors, selected by a hierarchal visual attention process and guided and processed using task, scene, function and object contextual knowledge improve image-based recognition processes? This is clearly addressing issues central to the cognitive vision approach. The two applications that the project will address are:
One post-doctoral researcher will be hired for this contract and a new PhD student will be in place.
The researcher hired will be responsible for undertaking research in the areas listed above, as well as cooperating with other research staff and students working on the project. The specific scientific work will depend on the project schedule, the progress of other members of the research group and abilities of the researcher.
Applicants for the post must have a PhD in an appropriate area, such as computer vision or image processing and should have experience with the MATLAB, C, C++ or JAVA programming languages and good mathematical skills.
The post is on the AR1 scale (18893-28279 pounds/annum). Placement for the post is according to experience and qualifications.
Other research duties will include involvement in the preparation and presentation of demonstrations, maintenance of the research group's publicity WWW pages, a stint at organising the Institute of Perception, Action and Behaviour's weekly seminar, visiting other vision research sites (mainly in the UK and Europe), the preparation and presentation of papers, attendance at conferences and internal seminars. The applicant will be expected to keep abreast of relevant international research in the area and to prepare and present conference and journal papers. The applicants should be able to produce evidence of the ability to conduct and publish original research.
The team member is also expected to contribute to the general project welfare, such as general software maintenance work, documentation, report writing, graphics coding, etc.
Departmental duties may include some teaching or tutorial work, such as presentation of internal demonstrations, second supervision of undergraduate, MSc and PhD students and development and marking of course work.
The researcher may be asked to attend selected MSc course modules to acquire familiarity with topics relevant to their work duties.
The researcher will work at the King's Buildings site, which houses the MVU Research Laboratory. The King's Buildings site is in a residential area about 3km south of the centre of Edinburgh, a pleasant and exciting city with considerable cultural attractions. Several recent surveys placed Edinburgh as one of the most desirable urban locations in the United Kingdom.
Normal Department working hours are 9:00 am to 5:15 pm, Monday to Friday inclusive, for a total of 35 hours per week. Holiday entitlement is six weeks per year, plus public holidays, to be taken during University vacation periods. An optional University Superannuation Scheme is available.
The post is available immediately and will last until September 30, 2007.
Applications should include a curriculum vitae (3 copies) and the names and addresses of two referees.
Further particulars are available from, and applications are sent to:
Irene MadisonThe closing date for applications is June 30, 2004.
In your application letter, please quote reference number 3002012.
The application form can be found at www.jobs.ed.ac.uk, which also allows an online application procedure.
NOTE: This statement does not of itself constitute a contract or conditions of service.
Back to main page.