A Psychovisual Threshold for Generating Quantization Process in Tchebichef Moment Image Compression

Abstract

A human visual system can hardly respond to small differences in image signals. A full colour image carries a certain amount of perceptual redundancy for the human eyes. The sensitivity human eye of the color image can be measured by a psychovisual threshold. The sensitivity of the human eye is useful for perceptual visual image in image compression. The quantization tables are obtained to determine psychovisual threshold that can be perceived visually significant by the human eye. This paper introduces the concept of psychovisual threshold into Tchebichef moment image compression. This paper will investigate the contribution of each moment coefficient to the image reconstruction. The error threshold from the contribution of its moments in image reconstruction will be the primitive of psychovisual threshold to an image. This paper presents a new technique to generate quantization table for an optimal TMT image compression based on psychovisual error threshold. The experimental results show that these new finer quantization tables provide a statistically better image quality output at lower average bit length of Huffman's code than previously proposed TMT quantization.