The project investigated 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 two applications to be investigated are:
1.Detecting, understanding and discriminating between similar types of interactions, such as two people fighting versus meeting and greeting.
2.Analysing crowd scenes, where tracking of individuals is only possible over short time periods. 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).
The project showed it was possible to detect dangerous situations through the use of video data, such as dangerous overcrowding in crowds and fighting in street scenes.
We successfully modeled crowd local disturbances and crowd blockages using a combination of optical flow and statistical modeling.
Fighting detection was also successful, based on a combination of person tracking and statistical modeling of properties of interactions.
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
