Font Size: a A A

Steganalysis Of Spread Spectrum Image Steganography

Posted on:2008-10-24Degree:MasterType:Thesis
Country:ChinaCandidate:L P ZhangFull Text:PDF
GTID:2178360215479844Subject:Communication and Information System
Abstract/Summary:PDF Full Text Request
After The September 11, 2001, attacks,steganography and steganalysis,as a new domain of Information Hiding, become a hot focus by more and more people. Steganography embeds messages in multifarious cover signal (words, image, voice et al), whose aims is to escape attacker's suspicion and detecting. Steganalysis, adverse side of steganography, detect hiding messages presence and extract message.Steganalysis has three sides: detecting presence /absence of hidden message, estimate location(s) of the hidden message or estimate the embedded message length, extract the hidden message. And now, many papers and monographs on steganalysis emphasize particularly on the first side. This paper researches all three sides as a synthesis of steganalysis. We focus on popular spread spectrum steganography and try to put emphasis from detecting presence on estimating location of hidden message.This paper firstly presents creative frame on detecting presence of hidden message. And then we present an improved steganalysis algorithm based on Block-DCT steganography. In order to improve the detecting ability, we add absolute value and 8-domain to pairs of neighboring pixel of inside block and outside block. Though this method we can broaden the difference of these two distributions in steg image, which can make detecting more correctly.This paper secondly presents creative frame on estimating location of hidden message. We present a new steganalysis method of sequential steganography. If cover signal is stationary, both sequential probability ratio test (SPRT) and optimal SPRT are used to estimate message beginning position and ending position. Still farther, when taken into account total error, we present a better k0 and k1 equations. If cover signal is non-stationary, we present a new method: though windows methods'to calculate every sequential variance, then threshold the SPRT and process CUSUM. So we can improve the accuracy of SPRT and can detecting steg image of low embed alpha.At last, we also program a simple detecting system. This makes steganalysis not a single aims but a system.
Keywords/Search Tags:Steganalysis, Steganography, Spread Spectrum, Information Hiding
PDF Full Text Request
Related items