Fault Tolerance in Small World Cellular Neural Networks for Image Processing

In this paper, a fault tolerant CNN (cellular neural network) using a small world network connection for image processing in embedded systems is described. In embedded systems, there are problems such as compactness, power consumption, reliability and speed, so new processor architecture is required. In this paper, we apply a small world network connection to cellular neural networks. The small world network features small links to the destination from the place of origin using local connection information of a random nature. The concept is derived from sociology. We then provide neural image processing algorithms suitable for CNN with the small world network model. Special attention has been given to typical neural algorithms for image processing, such as noise reduction and edge extraction. Further, we are proposing a fault tolerant architecture for the CNN, using majority voting.