As the digital age is becoming an unavoidable part of our daily life people are becoming more concern about their privacy and security of their digital communications. Steganography and cryptography are the two most frequently used merits of digital life to fulfill these needs. In this study, a new spatial domain chaotic steganography scheme is presented utilizing also new fractal stream encryption algorithm to hide Huffman encoded and compressed Turkish texts. The study composes of four main phases. Firstly, a sample of Turkish newspaper columnist corpus is gathered to obtain not only the frequencies of letters including special Turkish characters but also punctuations, spaces, newlines, quotations, etc. As a result of the first phase, a static Huffman encoding dictionary is obtained for 102 encountered characters. Secondly, super Mandelbrot sets are used with the Logistic map to obtain one-time-pad stream keys to encrypt compressed texts. Thirdly, the LSB (Least Significant Bit) plain of the cover image is analyzed morphologically to find low entropy pixel locations so that in steganography phase spotted locations can be avoided and scheme can become more resistant against stego analysis. Lastly, the chaotic steganography phase is the final phase in which data hiding pixel is selected according to logistic map chaotic function then LSB of the selected pixel is updated. Many tests are carried out on many images to show the image quality namely PSNR, SNR, SSMI, UIQI, Entropy, MSE, and histogram. Results indicate that proposed schemes are successful for encryption and offer robust steganography.
Read full abstract