Lattice vector quantization with direct indexing as data differencing technique

A huge range of approaches to deal with the task of image coding with lossy compression based on wavelet and vector quantization has been developed. Lattice Vector Quantization (LVQ), a variant of vector quantization is a lossy image compression technique. In this paper LVQ is applied on the transformed image as lossy and data differencing technique. While used as data differencing technique, the indices of quantized vectors with the indices of error vectors are sent to the decoder. To achieve this, extra computations are required at the encoder. Indices for lattice vectors are generated by proposed direct indexing method which is independent of the indexing of leaders and enumeration of the lattice points. PSNR values and visual performance of various test images are compared in both cases.