Novel Image Compression using Discrete Balanced Multiwavelet Transform and the Minimum List SPECK (MLS) Embedded Quantization Algorithm

Abstract

This paper presents a study of image compression carried out by adapting the embedded coding techniques to the discrete balanced multiwavelet transform (DBMWT). Balanced multiwavelet transform overcomes the lack of high approximation orders of traditional multiwavelet bases. Embedded coding techniques offer primary advantages of progressive transmission and embedded code file compared to traditional coders. Also the method to adapt the subband structure of a multiwavelet transformed image to embedded coding techniques that exploit the 4 subband structure of wavelet transform has been discussed. A modified version of the SPECK algorithm has been implemented along with Discrete Balanced Multiwavelet Transform. Finally image compression performance of the proposed algorithm has been evaluated for various test images and found to outperform existing compression schemes.