METHOD OF DETECTION INTEGRITY VIOLATIONS OF DIGITAL IMAGE BY PERLIN NOISE

Abstract

Context. Today informatization is one of the priority directions of economic sectors development. Introduction process of new information technologies in all society spheres is impossible without solving the problem of information security. An integral part of which is the development of methods for determining the authenticity of digital images, the construction of which in many areas is crucial. All thismakes the topic of this work is extremely important. Objective. To achieve the goal the following tasks are solved. Performed the research of digital image parameters, the presence or absence of the characteristic features of which allows to establish integrity violations of the digital image for the present forgery type. It is proposed to use for forgery detection ensuing form of singular values, as a digital image parameters, that uniquely define a matrix of digital image – a discrete function of the speed change ratio of block singular values. Based on developed form digital image parameters, was architected a way of increasing the efficiency of forgery detection method. Method. This paper presents a method for forgery detection of the digital image, developed on the basis of a common approach, which is use matrix analysis, perturbation theory, and provides an opportunity to solve the problem of detection integrity violations of the digital signal. Under the digital image forgery in the work refers to a integrity violation of the digital image, which is based on realistic visual simulation of natural phenomena, clouds and sky, as a result of Perlin noise generation. Results. In order to test the effectiveness of the modified forgery detection method, computational experiments performed for digital images that are stored without losses, to determine errors of the first and second type. Conclusions. The scientific novelty of the work is as follows: presented method of increasing the efficiency of the method of falsification detection, resulting in the generation of Perlin noise, based on the application instead of a discrete function of singular values of the block, a discrete function rate of change of the relationship matrix of singular values of image block. The result of work is a program product for detecting violation of the integrity of the digital image, based on Perlin noise, which can be used to supplement the complex detection systems integrity violation of the digital image