Performance Analysis of Stego-Image Security in Machine Learning

A numerical way to deal with steganalysis is introduced in this paper with straight steganography being the primary core interest. A numerically formal meaning of steganalysis is given trailed by definitions for uninvolved and dynamic steganalysis. The steganalysis issue is figured as visually impaired framework ID and conditions for recognizability (effective steganalysis) are determined. A methodology to deliberately misuse any accessible spatial and worldly variety data for proficient steganalysis is additionally examined. Test results are given for steganalysis of Gaussian circulated, spread range picture steganography and watermarking. The proposed procedure is seen to deliver great outcomes for an assortment of execution measures.