Dictionary Learning in Reduced Dimension Space for Image Steganalysis
Guoming Chen, Qiang Chen, Dong Zhang · 2014
In this paper we proposed a dictionary learning and dimensionality reduction (DLDR) scheme for image steganalysis. We construct a structural discriminative dictionary which is learned from the reduced dimension space and exploit the discriminative information in stego-images. Simulation results verify the effectiveness of the proposed approach and the performance is considerable. Both the dictionary and sparse coding can be correctly computed and the learned dictionary using the proposed method can be used to improve image steganlysis.