Edge based image steganography with variable threshold

Abstract

With the rapid escalation of the usage of sensitive data exchange through internet, information security has become a vulnerable realm of concern. Numerous security policies are incorporated, steganography being one of them. In this paper, a new image steganography technique facilitating embedding and extraction mechanisms have been proposed. The classic technique of edge detection is employed in which the threshold is made to vary for every new input test case. Relying on the threshold generation methodology, each edge image is generated in accordance with a distinct key threshold value. This in turn directs the insertion and extraction strategies. By using different performance metrics, like Payload, Mean Squared Error, Peak Signal to Noise Ratio, Structural Similarity Index, Kullback-Leibler Divergence, Cross-correlation, etc., this procedure substantiates better results. It has quite a remarkable embedding capacity (i.e. Payload). Also the values of Peak Signal to Noise Ratio and Structural Similarity Index indicate that the imperceptibility of the stego-image is well-maintained. The statistical results obtained reassure the undetectability of the hidden image. This approach may be used to deliver a status of security to a system which communicates secure data files through any public surveillance medium.