Image Compression Using Grey-Based Neural Networks In the Wavelet Transform Domain

Abstract

The wavelet transform has recently emerged as a powerful tool for image compression. In this paper, the Grey theory is applied to a two-layer modified Competitive Learning Network (GCLN) to generate optimal solutions for VQ. In accordance with the degree of similarity measures between training vectors and codevectors, the grey relational analysis is used to measure the relationship degree among them. The input image is first decomposed into four subbands in the 1-level wavelet transform. Then the corresponding transformed coefficients are trained using GCLN to form individual codebooks for each subband. The compression performances using the proposed approach are compared with GCLN and the conventional vector quantization LBG method. Experimental results show that promising performance can be obtained using the GCLN with wavelet decomposition.