Information Security is becoming an extremely important part of Data Communication. Steganography plays an important role in writing hidden messages in such a way that no one apart from the sender and intended recipient even realizes there is a hidden message. On the other hand, Steganalysis is the art and science of detecting embedded message based on visual inspection, statistical analysis or other methods. Most of the current steganography techniques that hide secret messages with suitable lengths can be detected by steganalysis techniques. In this thesis we propose a new steganography technique with a goal of minimum detectability for the hidden message. Towards that goal, the secret message is subjected to two phases of compression before embedding: in the first phase,the message is first compressed by removing the weak words and replacing some expressions with their commonly used abbreviations. In the second phase, the resultant compressed message is further compressed using the Huffman lossless compression technique. Finally, the compressed secret message is embedded into the cover image based on the modulus three of the difference between DCT coefficients of the cover image during JPEG compression process.
The proposed technique is tested using seven different standard test images that are commonly used in image processing researches, in addition to five different secret messages with different lengths collected from random paragraphs. The experimental results of the proposed algorithm have been evaluated based on different benchmarks techniques such as Shannon’s entropy, AAD (Average Absolute Difference), MSE (Mean Square Error), SNR (Signal to Noise Ration) and PSNR (Peak Signal to Noise Ratio). The results prove that the proposed technique outperforms the LSB technique in hiding secret messages with different lengths inside cover images.