Advanced methods of detection of the steganography content

Abstract

In this paper, we deal with the classification of the steganography content. Some illegal activities can perform steganography for stealing the secret information from a company internal network. Therefore, we must be prepared to protect our data. To detect steganography content, we have counter-technique known as steganalysis. There are different types of steganalysis, based on the existence of the original artifact (cover work) or if we know which algorithm embed a secret message. For practical use, most important are methods of blind steganalysis, that can be applied to the most compact and ordinary cover work - JPEG image files. This paper describes the methodology to the issues of JPEG image steganalysis. It is crucial to understand the behavior of the targeted steganography algorithm. Then we can use it is weaknesses to increase the detection capability and success of classification. We are primarily focusing on breaking the DCT steganography algorithm OutGuess2.0 and secondary on breaking the F5 algorithm. We are analyzing the ability of the detector, which utilizes the calibration process, blockiness calculation, and shallow neural network to identify the presence of steganography message in the suspected image. This approach is an improvement over our previous researches. Contribution and new results are discussed.