engleski

A Multi-Agent-Based Approach to Face Detection and Localization

This paper presents a multi-agent-based approach to face detection and localization in colour images. In the initial step of the process the autonomous behaviour agents are distributed throughout the image. In subsequent steps the agents detect and mark skin-colour pixels using the HSI model and the innate agent-s behaviour functions, such as diffusion, self-reproduction and dying. During the final step, the results of all the families of agents are combined and the face-like regions are detected. After face-like region localization, the shape information is used to decide if the region really represents a human face. The approach was tested using both the XM2VTS database of frontal face images and images containing people in natural situations.

multi-agent system; behaviour-based agents; image segmentation; face detection; face localization

nije evidentirano