A framework for steganalysis of images in internet traffic

This paper presents a framework for monitoring internet traffic to detect the presence of hidden information in images that were subjected to steganography. The framework of steganalysis environment comprises three stages: pre-processing, statistical and distortion. The environment enables efficient steganalysis of internet traffic and gives an indication of the robustness of the various steganography techniques used in the study. An extensive set of cover images was used to set the detection thresholds. The impact of distortion, applied to destroy hidden information, was also assessed by calculating the cover image signal to noise ratio. The effectiveness of the complete environment as an integrated steganalysis system was evaluated by injecting a wide variety of clean as well as stegoed images that were not used during threshold setting. The results indicate that the integrated environment achieves a high detection rate with acceptable levels of false positive and false negative detections.