Color Image Compression Using Spread Grey-Based Neural Networks in the Transform Domain

In this paper, a new Grey-based Competitive Learning Network (GCLN) for Vector Quantization (VQ) and Spread GCLN (SGCLN) for color image compression in the Discrete Cosine Transform (DCT) and Mean value / Difference value Transform (MDT) domains are proposed. A spread-unsupervised scheme based on the competitive learning neural network using the grey theory is proposed so that on-line learning and parallel implementation is feasible. In the GCLN, the grey theory is applied to a two-layer modified competitive learning network (MCLN) in order to generate optimal solution for VQ. In accordance with the degree of similarity measure between training vectors and codevectors, the grey relational analysis is used to measure the relationship degree among them. The color image information transformed by DCT or 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 SGCLN method to generate the VQ codebook. The experimental results show that promising codebooks can be obtained using the proposed GCLN and SGCLN for color image compression in the transform domains.