Embedded zerotrees wavelet image coding using source polar codes

Abstract

Following the directions of Shannon's duality of source coding and channel coding, source polar codes have been found and proven to be a rate distortion achievable lossy source codes. Nowadays, research of these codes which are developed from channel codes is still in the stage of theoretical simulation, and there is not specific application being found. Based on the excellent rate distortion performance of source polar codes, we propose a novel embedded zerotrees wavelet image coding method using source polar codes creatively in this paper, which is called PC-EZW (Polar Coded — Embedded Zerotrees Wavelet). PC — EZW could realize a rate-distortion controllable lossy image compression, and the experiments show that PC-EZW has higher coding speed, less storage budget and scan times than EZW (Embedded Zerotrees Wavelet) in some standard test images. At the same time, PC-EZW offers the possibility of parallel process of lowpass subband and highpass subband.