PARITY AND PIXEL VALUE DIFFERENCE BASED IMAGE STEGANOGRAPHY ALGORITHM

摘要

In the era of digitalization, most of the data is transmitted through the internet. Digital data security is a prominent issue for this century. There are the variety of techniques used to protect data during communication. Among all of them, apple of my eyes steganography method plays a significant role in digital coinmunication. Steganography method reduces chance of information detection while data transmission. Steganography is an art of hiding the existence of a message between the sender and intended beneficiary. Steganography has been used to conceal secret information in various types of files, such as digital images, audio and video. The three most significant parameters for steganography are imperceptibility. payload, and robustness. Different applications have diverse requirements of the steganography method. Unique design of image steganography algorithm based on pixel parity and pixel value difference for gray scale and color image has been presented. It has numerous advantages compared to other tech¬niques. Moreover, it can withstand with all kind passive steganalysis and discuss a steganography capacity quantify in the presence of passive steganalysis. For steganalysis, Machine learning technique is used. The machine is trained by large set of data and based on the trained data, another set of cover and stego images are used for estimation and detection of those images. Classifier is providing output in from of confusion matrix and that is used to plot Receiver Operating Characteristic (ROC) curve. Furthermore related to analysis, carrier image quality assessment after embedding information is highly significant. The subjective evaluation of sensitivity to distortions due to data hiding is investigated via. visual attack and bit plane attack. Ultimately, results demonstrate that presented system successfully hides a significant amount of payload within digital images without being detectable by an observer.