In this paper two novel steganalysis methods is presented based on co-occurrence matrix of an image. It is shown that by using features extracted from this matrix, we can differentiate between cover and stego images. These features include energy, entropy, contrast, inverse difference moment, maximum probability and correlation. We use SVM classification for separation of cover and stego images. In the second method, we directly use the diagonal elements of co-occurrence matrix as a feature and we remove some bit planes of an image for feature dimension reduction. Experiments on the 1800 images for LSB and LSB matching (LSBM) algorithms are performed with different rates for the same database. By comparing the results, we also show that the detection rates effectively are increased with respect to previous steganalysis techniques.
Rights and permissions | |
This Journal is an open access Journal Licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. (CC BY NC 4.0) |