Image Compression Using Mean-Removed And Multistage Vector Quantization In Wavelet Domain

Abstract—Abstract The main purpose of the image compression is to reduce the size of the image with
minimum degradation. Internet world use data, images and videos massively. The images are compressed with the higher compression ratios using lossy compression techniques for transmission and efficient storage of the images. The quality of the image degrades in these techniques. There are different types of technique available for the lossy image compression. Lossy image compression technique presented in this paper is easy to implement, secure and efficient which is called Vector Quantization (VQ) based on wavelet transform. In this paper design and implementation of two types of vector quantization techniques Mean-removed vector quantization, Multistage vector quantization (MSVQ) using wavelet transforms are proposed and the performance is compared or quality of the reconstructed image on the basis of quality measures. Major artifacts like blockings and loss of edge information are observed in VQ based image compression technique. The use of wavelet with the VQ technique shows improvement in this area. The performance is analyzed using special edge based quality measures MESSIM. Mean edge based similarity index. The performance of the multistage vector quantization is much better than the mean removed vector quantization in terms of PSNR and MESSIM.