Audio Stego Intrusion Detection System through Hybrid Neural Tree Model

Abstract

Steganography and steganalysis in audio covers are significant research topics since audio data is becoming an appropriate cover to hide comprehensive documents or confidential data. This article proposes a hybrid neural tree model to enhance the performance of the AQM steganalyser. Practically, false negative errors are more expensive than the false positive errors, since they cause a greater loss to organizations. The proposed neural model is operating with the cost ratio of false negative errors to false positive errors of the steganalyser as the activation function. Empirical results show that the evolutionary neural tree model designed based on the asymmetric costs of false negative and false positive errors proves to be more effective and provides higher accuracy than the basic AQM steganalyser.