Mikołaj Płachta
supervisor: Artur Janicki
Information security is becoming an increasingly important issue in our world. Therefore, more and more sophisticated attack methods are emerging, among them the methods using digital image steganography. Therefore, in the first part of my research for my doctoral thesis, I focused on the methods of automatic detection of digital steganography. The main part of my research is the use of machine learning methods, in particular deep neural networks. For data discovery methods, a couple of methods were tested. The first was the analysis of the discrete cosine transform and an attempt to teach a neural network to classify JPG images. The next steps were to use different coefficients that can extract the relevant information from the pictures and help to make the detection. For this, we used DCT Residuals (DCTR) and Gabor Filter Residuals (GFR). The set of BossBase, known in the world of steganography, was used for the tests. It consists of carefully selected black and white images in the amount of 10,000. On its basis, files with encrypted data were generated using the jUniward and nsF5 algorithms. Pairs of pictures (without steganography and with hidden data) were used to train and verify the network. The obtained results and conclusions will be presented during the presentation.