Understanding far and close proximity human-human interaction observed from a distance is a necessary step towards automated suspicious or antisocial behaviour detection. Most previous work on human-human interaction has made the implicit assumption that interactions occur only at immediate spatial and temporal proximity between the subjects concerned. We propose a more realistic application of human-human interaction detection from surveillance data where the subjects of interest tend to be represented by few pixels relative to the rest of the scene. The subjects are represented by relatively few pixels since surveillance cameras are usually placed to maximise area coverage therefore there is a significant distance between the camera and the physical scene. This in itself is not so much of a disadvantage when we consider that interactions among subjects can occur between quite large distances in space. Our technique uses a spatial and temporal saliency measure to extract and select features using modifications to Kadir and Brady's scale saliency and Hung and Gong's temporal saliency algorithms respectively. From this, a hierarchical multi-scale model of a single person, his/her body pose and groups of people is formed. A person is represented by an elliptic blob where prominent oval-shaped parts are formed into a configuration. Interactions are identified by finding temporally correlated salient changes (we call events) in the probability distributions of our multi-scale configuration model. In this paper we will show how pose or configuration based models of the human body can provide a rich framework for modelling human-human interactive body behaviour even when body parts are occluded. In particular, the framework is suitable for extracting salient features from the human body where each part is represented by a few pixels in each image frame. The work is highly relevant to the development of automated systems for suspicious and antisocial behaviour detection and prevention.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.