Self-organizing neural network domain classification for fractal image coding

This paper presents a scheme for improving encoding times for fractal image compression. The approach combines feature extraction with domain classification using a self-organizing neural network. Feature extraction reduces the dimensionality of the problem and enables the neural network to be trained on an image separate from the test image. The self-organizing network introduces a neighborhood topology for classification, and also eliminates the need to specify a priori a set of appropriate image classes. The network organizes itself according to the distribution of the image features observed during training. The paper presents results showing that this classification approach can reduce encoding times by two orders of magnitude, while maintaining comparable accuracy and compression performance.