CAVERNOUS ERUDITION HIERARCHICAL REPRESENTATIONS FOR IMAGE STEGANALYSIS

Abstract

The prevailing detectors of Steganography communication in digital images mainly consist of three steps. Residual computation, feature extraction and binary classification. The alternative approach to Steganalysis using digital images based on alternative approach to Steganalysis using digital image based on Convolutional neural network (CNN). The proposed CNN has a different structure from the ones used in conventional computer vision tasks (CVs).this to replicate and optimize these key steps in unified framework and learns hierarchical representation from raw images. Steganalysis using three state of the art steganographic algorithms in spatial domain e.g. HILL is better than WOW and S-UNIWARD. Selection channel aware [SCA TLU-CNN] overcome the TLU-CNN methods.