A NEW GENETIC ALGORITHM AND BLOCK-BASED METHOD FOR IMAGE STEGANOGRAPHY


Res. Asst. ZEYNEP SARI

Thesis Type: Postgraduate

Institution Of The Thesis: Kocaeli University, Teknoloji Fakültesi, Bilişim Sistemleri Mühendisliği, Turkey

Thesis Supervisor: Mehmet Yıldırım

Approval Date: 2022

Thesis Language: Turkish

Abstract:

Data security is an important issue in all types of communication, especially in confidential data communications. Encryption of data, that is, cryptology, is one of the classical methods to ensure the security of data. Encryption of data keeps the data content confidential, but since the existence of the data is not confidential, it can be observed or damaged by undesired persons. Steganography is the science that allows the existence of data to be hidden. Steganographic studies use media called carriers or covers to hide the existence of data. Carrier media; can be video, audio, image, or program files. The media obtained by hiding data to the carrier media is called stego media. In spatial domain image steganography, the carrier media is digital image files. Data hiding operations are performed directly on the pixel values of the carrier image. Many researchers have developed techniques in this field and suggested techniques. In this thesis, a new steganography technique was developed based on the least significant bit displacement technique used in spatial domain image steganography. The proposed technique uses the techniques of separating the carrier image into blocks and channel selection, which are also available in the literature, for the secure transmission of data. The security of the data, which is one of the important and basic criteria of steganography, has been tried to be increased by selecting the channel with the genetic algorithm in the block data and the security of the data has been ensured by the key generated by the genetic algorithm. In practice, grayscale image files of five different sizes were used as confidential data in order to hide data of different capacities on a fixed-size grayscale carrier image. Instead of sequentially selecting the carrier image pixels in the classical least significant bit technique, this selection is made with a genetic algorithm, preventing the data from being easily retrieved without a key. As a result of the application, it has been observed that both the security of confidential data can be increased compared to the techniques in the literature and the presence of the data can be better hidden by reducing the deterioration in the carrier image.