CAVIAR: Context Aware Vision using Image-based Active Recognition

BRIEF PROJECT OVERVIEW

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:

1. City centre surveillance: Many large cities have nighttime crime and antisocial behaviour problems, such as drunkenness, fights, vandalism, breaking and entering shop windows, etc. Often these cities have video cameras already installed, but what is lacking is a semi-automatic analysis of the video stream. Such analysis could detect unusual events, such as patterns of running people, converging people, or stationary people, and then alert human security staff.
2. Marketers are interested in the behaviour of potential customers in a commercial setting, such as what sequence of locations do they visit, how long they stop at particular locations, what behavioural options do typical customers take, etc. Automatic analysis of customer behaviour could enable evaluation of shop layouts, changing displays and the effect of promotional materials.

A more detailed summary of the project objectives.

The project started on October 1, 2002 and continues until September 30, 2005.

PROJECT ACTIVITIES

• Foveated and other feature extraction and grouping,
• Integrating feature, object and top-down priming for spatial and temporal attention,
• Representing and recognising objects, contexts and situations,
• Learning instances of the representations from visual evidence and
• Reactive and top-down control of the recognition process.

Project activities, achievements, and news.

RESEARCH RESULTS

PPT/PDF slides from the final project presentation/review:

1. Overview, all workpackages except for WP1 and WP5, administrative details (PDF)
2. WP1: Active image acquisition and description (PDF)
3. WP4+5: Acquiring and using models for contexts, situations and control of perception (PDF)

You can also find some:

1. Publically downloadable benchmark datasets
2. Project publications
3. Results examples
4. Performance statistics

WORKSHOPS ORGANISED

• HAREM05: Human Activity Modelling and Recognition, Oxford, UK, Sept 9, 2005.
• PETS04: Sixth IEEE International Workshop on Performance Evaluation of Tracking and Surveillance, Prague, Czech Republic, May 10, 2004.
• VSCA03: Computer Vision System Control Architectures, Graz, Austria, March 31, 2003.

PARTNERS

The CAVIAR team leaders are:

• Prof. Robert Fisher (coordinator), Univ. of Edinburgh (United Kingdom)
• Dr. Jose Santos-Victor, Instituto Superior Tecnico, Lisbon, Portugal
• Prof. James Crowley, Institut National Polytechnique de Grenoble, France

More information about the scientists working on the project.

RESEARCH POSITIONS AVAILABLE

At the moment, there are no open posts.

COORDINATOR