Cellular neural network based VLSI architecture for image processing

The paper presents a way for increasing a size of images which can be processed using Cellular Neural Networks based VLSI circuits. Basic idea is to reduce a number of rows which are being simultaneously processed in a network. To verify the proposed ideal a VLSI integrated circuit which allows for realizing selected gray-level image analyses, has been designed, manufactured and its performance was examined. In addition, a possibility of applying the proposed concept in realizing image processing operations where global signal propagation is crucial such as e.g. halftoning, have been discussed.