COMPARATIVE RESEARCH OF ORTHOGONAL TRANSFORMS FOR COMPRESSION OF DIGITAL IMAGES

Abstract

In the paper the comparative research of the most common orthogonal transformations for compressing digital images has been conducted. This article substantiates scientific and practical value and prospects of using orthogonal transformations. Karhunen–Loeve, Haar, Walsh-Hadamard transforms, discrete and cosine Hartley transforms, their essential advantages, disadvantages and features for application in the methods and algorithms for image compression have been analyzed. In addition, transformant values of transformation of a digital image unit have been examined and evaluated during scientific research. The results of study and comparative analysis of indicated orthogonal transforms prove that discrete Hartley transform compared to other orthogonal transforms has significant advantages for practical realization and has a smooth transformant structure of the digital image unit providing high compressive properties. Consequently, it is highly reasonable to use discrete Hartley transform for efficient compressing digital images.