Lossy Image Compression Using Karhunen-Loeve Transform Based Methods

In this paper, we discuss image compression techniques based on the eigenvector matrices used Karhunen-Loeve Transform (KLT) is obtained. Two novel methods are proposed for grouping of eigenvectors via vector quantization in KLT subspace. Various codebook sizes tested purposes. The first approach uses autocorrelation geometrically clustering into fewer numbers vectors. approach, performed using principal component directions matrices. second has according to their usage frequencies. qualities reconstructed test images compared with DCT JPEG and Wavelet JPEG2000 PSNR metric. Experimental results show that methods, particularly method, give plausible competitive results.