A New Variation of Singular Value Decomposition

This paper presents a new variation of Singular Value Decomposition (SVD) computation for image compression. It modifies the SVD operation with two additional pre-processing steps, i.e. Pixel Interleaving and Linearization process. An input image is firstly scrambled by Pixel Interleaving operation. This interleaved image is further linearized before applying SVD calculation. As documented in the Experimental Results Section, this simple approach yields a promising result in the image compression task compared to that of the classical SVD-based technique.