Reactive agents for adaptive image analysis

In this paper, we describe a novel computational method for image segmentation that utilizes a collection of distributed autonomous agents. The goal of the agent-based image segmentation is to efficiently locate and label various homogeneous regions from a digital image. To do so, the autonomous agents, being distributed computational entities, will operate at the individual pixels of the given image, and execute a number of reactive behavioral responses. The proposed method of image segmentation with the behavior-based reactive agents can exhibit several distinct features; namely, the computation is adaptive with respect to the locality of the given image and hence efficient in locating a homogeneous region, reliable as the agents proceed simultaneously from different image locations, less sensitive to the noise in the given image as the agents will disperse gracefully when they are moving away from the homogeneous region, and easy to represent and implement.