Publication

Abstract : With the increase in the amount of data transferred online, there is an increasing requirement of security. Steganography is one field of data encryption which focuses on hiding a secret message inside a cover image. A normal person intercepting the image would not be able to identify the secret message. This paper proposes a new algorithm which performs a variation of LSB substitution method in the spatial domain, where the cover image is split into various sub-images and data embedding is done on the sub-images individually. Two LSB embedding methods which operate using three threshold values are also proposed and a key matrix identifies which one of the two methods is to be applied to each of the sub-images. The results and analysis show that the proposed technique is more secure with high PSNR values and is resistant to steganalysis. It is also found that the proposed algorithm has high data embedding capacity.