Steganography detection of whole-image-DCT-based watermarking in gray-scale images
Document Type
Article
Publication Date
12-1-2003
Abstract
In this paper, the recently introduced steganography detection system (SDS), is applied to the detection of whole image DCT-based steganography, in gray-scale images. Here, the watermarking is applied in the first 1000 transform coefficients of the entire image. The normalized differences in the coefficients of the DCT transforms of the watermarked and unwatermarked images from the original are treated as features. SDS utilizes statistical preprocessing, over subsets of the set of transform coefficient values of each image, organized as feature vectors. These vectors are then fed into a simple neural network classifier. For the experiments conducted here, using 1096 images, SDS achieves perfect detection rate with no misclassifications errors.
Identifier
4944258223 (Scopus)
ISBN
[9608052874, 9789608052871]
Publication Title
Recent Advances in Intelligent Systems and Signal Processing
First Page
162
Last Page
166
Recommended Citation
Song, Sui; Zou, Dekun; Manikopoulos, Constantine; and Shi, Yun Qing, "Steganography detection of whole-image-DCT-based watermarking in gray-scale images" (2003). Faculty Publications. 13831.
https://digitalcommons.njit.edu/fac_pubs/13831
