Image stenography with improved PSNR

Abstract

Communicating secretly has always remained an interesting and important research field. Commonly used method to safely share information is cryptography. However cryptographic methods are inherently vulnerable in a sense that cryptographic communication raises suspicion and is a recipe to break into which a determined analyst will always achieve. However unlike cryptography, steganography conceals the information. This is done to shed away possible chances of suspicion that something important is being shared covertly. Besides watermarking and fingerprinting, covert communication methods fall under the broader class of hidden communication. The hidden communication though has deep roots in history; yet modern communication systems in the form of digital media provide the best concealing mediums to achieve such secrecy. Images are the most favored cover objects which have extensively been used for hidden communication due to their wide circulation on the internet and inherent exploitable redundancies. Three most important properties of any steganography scheme are its capacity, quality and security. Any good steganography approach should maintain a delicate balance amongst these conflicting properties. Steganography capacity is mostly influenced by cover image size. However there is a tradeoff between steganographic capacity and stego image quality. This being a challenge is my research objective. To improve capacity while maintaining a good quality, a flexible LSB embedding method based upon correlation has been proposed. Objective analysis based upon PSNR and MSE has been used to measure the quality of steganography. The proposed method has been empirically validated and the results show that proposed method gives improved capacity, PSNR and security.