An Optimal Least Significant Bit Based Image Steganography Algorithm
Shuting Xu, Shuhua Lai · 2014
In this paper, we propose an effective Least Significant Bit (LSB) based steganography algorithm. The new algorithm is based on the classic K-means algorithm. We split bits of a secret message into clusters so that clusters of bits can be assigned to replace the LSB of each pixel of a cover image. To successfully use K-means we define a function to calculate the distance between the bits and the clusters. Bits can be moved among neighboring clusters based on the distance to the centroids of clusters. Since the classic K-means algorithm converges to an optimum, our approach leads to an optimized stego-image, compared to results of other LSB based approaches. Real test cases show that this approach can hide 60% of the size of the cover image without any noticeable visual artifacts.