#4

Permanent URI for this community

https://elar.naiau.kiev.ua/handle/123456789/25531

Browse

Browsing #4 by Author "Bobok, I.I."

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Розробка методу відокремлення цифрового зображення, збереженого у форматі без втрат, від зображення, збереженого із втратами. DEVELOPMENT OF THE METHOD OF SEPARATING DIGITAL IMAGE, STORED IN A LOSSLESS FORMAT, FROM STORED WITH LOSSES
    (2018) Бобок, І.І.; Bobok, I.I.
    Information integrity is one of the main properties of the information security object. Unauthorized actions on an image, using it as a container in a stenographical system, often lead to a difference in the format of the original image from the one whose integrity is violated. The establishment of the fact of such a format change is a pointer to the transformation of the image, to the violation of its integrity. The paper developed a method of separating a digital image, which was originally stored in a lossless format, from that which was saved to a lossless format from a losses format. The method is designed to determine the quality factor QF used in the primary compression. The basis of the method is the analysis of the normalized separation of the maximum singular numbers of blocks of the image matrix. The blocks are obtained as a result of the standard partition of the matrix. Concept of a pronounced local maximum of a function allowed automated the process of determining the quality factor QF of the primary compression of a digital image. Two algorithmic implementations of the method are proposed, which have high efficiency, which is determined by errors of the first kind. For one of the algorithmic implementations, errors of the first kind were not identified during the computational experiment. The average value of the number of errors in determining the quality factor QF of the primary compression of the original losses image was 7.5 % when using the first method for determining the pronounced local maximum and 6.5 % when using the second method. The computational complexity of the proposed algorithms of the developed method is determined by the number of blocks of the digital image that is subjected to expertise. For the image with the matrix became operations. If a color image is analyzed (for example, an RGB scheme), the proposed algorithms can be applied both to one (arbitrarily chosen) color component, and for each of them. The developed method is considered by the authors as the basis for developing a method for detecting violations of the integrity of a digital image under conditions of minor disturbances of the original image matrix.