Large-scale steganalysis using outlier detection method for image sharing application

Abstract

In past several years so many steganalysis techniques are emerged but none of them are not efficient for real world image sharing applications where millions of images are transmitted. In small scale steganalysis individual images are detected for suspicious payload. Here uses a different approach to determine most prolific steganographer who sends large volume of secret data in the network. A new technique is introduced to determine the steganographer out of large number of users, where each user transmits numerous images. In this method steganalytic features are extracted from image, distance between users are calculated and finding out the outlier user who deviate from the majority of other users.