A multi-agent system to segment living cells

This paper presents a multi-agent system for the segmentation of deformable moving objects. Our application concerns cell image sequences. The agents integrate three basic behaviors: segmentation, interaction and reproduction. The segmentation behavior is based on region growing methods. Pixels are evaluated upon static and motion-based criteria. The interaction behavior allows two agents to merge or to negotiate parts of regions. The negotiation can be seen us a segmentation refinement process done by the agents. Finally, the reproduction behavior defines an exploration strategy of the image frames. Agents can start other agents around them, or they can duplicate themselves in the next frame. Segmentation of the frames is done in pipeline, where previous information is used to help the process of the current frame. One unique agent model exists. Segmentation and reproduction behaviors are specialized on the execution time. Some results from an existing prototype are shown and analysed