CELLULAR NEURAL-NETWORK ARCHITECTURE FOR GIBBS RANDOM-FIELD-BASED IMAGE SEGMENTATION

We describe a novel cellular connectionist neural network model for th e implementation of clustering-based Bayesian image segmentation with Gibbs random-field spatial constraints. The success of this algorithm is largely due to the neighborhood constraints modeled by the Gibbs ra ndom field. However, the iterative enforcement of the neighborhood con straints involved in the Bayesian estimation would generally require t remendous computational power. Such computational requirement hinders the real-time application of the Bayesian image segmentation algorithm s. The cellular connectionist model proposed aims at implementing the Bayesian image segmentation with real-time processing potentials. With a cellular neural network architecture mapped onto the image spatial domain, the powerful Gibbs spatial constraints are realized through th e interactions among neurons connected through their spatial cellular layout. This network model is structurally similar to the conventional cellular network. However, in this new cellular model, the processing elements designed within the connectionist network are functionally m ore versatile to meet the challenging needs of Bayesian image segmenta tion based on the Gibbs random field. We prove that this cellular neur al network does converge to the desired steady state with a properly d esigned update scheme. Examples of CT volumetric medical image segment ation are presented to demonstrate the potential of this cellular neur al network fora specific image segmentation application. (C) 1998 SPIE and IS&T. [S1017-9909(98)00501-7].