Image Processing Algorithms for a Multiple-Valued Array Processor,

Abstract : A new digital image processor called multiple-valued array processor (MVAP) is effectively employed for systematic image processing without encoding and decoding because each pixel can be directly expressed by a single multiple-valued digit in the images with gray levels of several colors. In this paper, some properties of image processing in the MVAP are presented. Especially, the near-neighbor instructions can be attributed to template matching as the state transition function. In multiple-valued logic, there exist many templates because of its logical richness. The simplification of the state transition function is very useful for the effective execution of the image processing. The systematic design method of the image processing algorithm and its simplification using a minimization and its simplification using a minimization technique of multiple-valued logic functions are discussed. (Author)