An improved embedded image compression algorithm

This paper presents an improved image compression algorithm to make up a deficiency arising when the embedded zerotree wavelet encoding (EZW) is employed to uniformly scan the low and high frequency coefficients. According to the feature of the wavelet transform´s energy distribution, the scalar quantization is used to quantify sub-band coefficients of high frequency in order to reduce the IZ(Isolated Zero); the threshold sequence is then adjusted in the process of the successive approximation quantization to minimize distortion when recovering images in this paper. Simulation results show that this improved coding scheme is superior to the EZW algorithm in PSNR as well as image reconstruction quality.