IMPROVED BLOCK TRUNCATION CODING USING HOPFIELD NEURAL NETWORK

Block truncation coding (BTC), a recent technique used in the coding of image data, is based on the classification of pixels within a small image block into two classes. Earlier methods used statistical information for this classification. In the Letter, a new technique is introduced which uses a Hopfield neural network to define the pixel classes. Results are presented for four monochrome still images. The new algorithm is shown to provide improved performance when compared to the two previous BTC algorithms.