Vector quantization of images using modified adaptive resonance algorithm for hierarchical clustering

Most neural-network (NN) algorithms used for the purpose of vector quantization (VQ) focus on the mean squared error minimization within the reference- or code-vector space. This feature frequently causes increased entropy of the information contained in the quantizer (NN), leading to a number of disadvantages, including more apparent distortion and more demanding transmission. A modified adaptive resonance theory (ART2) learning algorithm, which we employ in this paper, belongs to the family of NN algorithms whose main goal is the discovery of input data clusters, without considering their actual size. This feature makes the modified ART2 algorithm very convenient for image compression tasks, particularly when dealing with images with large background areas containing few details. Moreover, due to the ability to produce hierarchical quantization (clustering), the modified ART2 algorithm is proven to significantly reduce the computation time required for coding, and therefore enhance the overall compression process. Examples of the results obtained are presented in the paper, suggesting the benefits of using this algorithm for the purpose of VQ, i.e., image compression, over the other NN learning algorithms.