Optimum Coefficients Of Discrete Orthogonal Tchebichef Moment Transform To Improve The Performance Of Image Compression

Abstract

This paper proposes an improved image compression scheme which utilizes optimum coefficients of discrete orthogonal Tchebichef moment transform. In the conventional moments-based-compression techniques, after making a trade-off between the conflicting factors, namely the quality of the reconstructed image and the compression ratio, the moments are chosen sequentially up to the desired order for reconstruction. But, as far as this proposed method is concerned, for reconstruction, it utilizes the optimally selected Tchebichef Moment Transform coefficients that yield better quality output image with a reasonable peak-signal-to-noise ratio for a desired value of compression ratio. Standard test images of different classes with various sizes (128x128, 256x256, 512x512 and 1024x1024) have been subjected to the proposed compression method for the block sizes 4x4 and 8x8 in order to assess its performance. The standard performance measures such as compression ratio, mean square error and peak-signal-to-noise ratio are considered in this study. The results reveal that the proposed method vis-à-vis the task of compression performs invariably well for the different classes of input images of above-mentioned sizes.