Text steganalysis using AdaBoost
Zhongliang Zhu · Journal of Communications · 2007
The statistical models and features of natural texts was analyzed,and it was pointed out that embedding mes-sages in texts will change the features of them.According to the changes,a blind detecting method was designed using AdaBoost.Five basic parameters of texts was extracted as distinguished feature vectors to discriminate natural texts and stego-texts effectively using AdaBoost.Experimental results show the high accuracy and reliability of the method.