Steganography App Based on Local Colour Statistics

In today’s world the usage of smart phones goes way beyond the common use. It can be used to send and receive secret messages concealed into different data transfer formats, like documents and images, that are invisible to human perception. This paper deals with the development of a Image Steganography Application for a smartphone based on statistical Gaussian Mixture Model of pixel representation by two colour model in the local neighborhood. Edges play important role in this statistical approach, since they are the regions of image high frequency that are capable of carrying payload without visual detection. The effectiveness of the proposed steganography technique was assesed by PSNR and L2 norm. The detailed description of an application developed using Matlab is provided as well.