Universal steganalysis using DWT and entropy features
Anupama K. Ingale, Nagaraj V. Dharwadkar, Pratik Kodulkar · 2016
Steganography is the process of embedding text information or image information into cover image. Universal Steganalysis detects the hidden information embedded in the digital object when the embedded algorithm is unknown. Universal steganalysis consists of two major parts viz., Feature extraction and Classifier selection and design. In this paper we propose a multilayer neural network for classification. Image is divided into micro blocks and feature extraction techniques like entropy and DWT(Discrete Wavelet Transform) are used to gather the information in both spatial and frequency domain which is then used for classification. The experimental results show that the output of cascaded neural network is 98%.