A spread neural network with fuzzy clustering technique applied to color image coding in the MDT domain

In this paper an unsupervised parallel approach called fuzzy competitive learning network (FCLN) for vector quantization (VQ) and spread FCLN (SFCLN) for color image compression in the mean value/difference value transform (MDT) domain are proposed. In the FCLN, the codebook design is conceptually considered as a clustering problem. Here, it is a kind of competitive learning network model imposed by the fuzzy clustering strategies working toward minimizing an objective function defined as the average distortion measure between any two training vectors within the same class. The color image information transformed by the MDT operation was separated into RGB 3-plane mean value and detail coefficients. Then the detail coefficients for each plane were trained using the proposed SFCLN method to generate the VQ codebook. The experimental results show that promising codebooks can be obtained using the proposed FCLN and SFCLN for color image compression in the MDT domain.