Parallel image processing shows certain general similarities to retinal processing and shares with optical computing the characteristics of parallelism. Simulation on digital serial computers (DSC's), although possible in the simpler cases, very soon becomes almost uncontrollable, requiring large memories and very long programming and execution times. Advances in present circuit technology now permit one to construct large arrays of interconnected parallel ogic elements resulting in real-time parallel machines where the action of a statement is simultaneous on all the points of the array. A typical processor of this kind is briefly described and an introduction to its code is given. Examples of symmetrical and directional functions and parallel algorithms are also presented. Possible applications of such processors in image processing, pattern recognition, and artificial intelligence are briefly discussed.

After an overview of the cellular logic image processor, CLIP 3, parallel algorithms for joining two points are described. These real time algorithms are simple, work properly, preserve the connectivity, are based only on logic, and show much of the behavior of processing occurring in sensory elements of animals.