Discrete wavelet transform based Hybrid image coding schemes for continuous tone images

Abstract

The thesis aims to propose new hybrid algorithms for wavelet newlinebased image compression of continuous tone still images to achieve low bit newlinerates The algorithms have been proposed by combining Discrete Wavelet newlineTransform Absolute Moment Block Truncation Coding Singular Value newlineDecomposition and Clustering of data for the purpose of vector quantization newlineThe proposed algorithms offer improved compression ratio The proposed newlinemethods utilize the property of wavelet transform in which most of the detail newlinesubband coefficients remain quantized to zero Also a set of Daubechies newlinewavelet templates are designed to be used with Cellular Neural Networks in newlinehigh speed image compression applications newlineIn this thesis two hybrid image compression algorithms that are newlinebased on Absolute Moment Block Truncation Coder AMBTC and DWT are newlineproposed The proposed method is very attractive due to its reduced newlinecomputational complexity newline newline