Cover Selection for More Secure Steganography
Reza Esfahani, Zynolabedin Norozi, Gholamreza Jandaghı · International Journal of Security and Its Applications · 2018
The lack of a cover bank with minimum detectability against various steganalysis methods is one of the major concerns in the field of secure communication.This paper introduces a method for preparing cover images with appropriate security.To this end, some important features of the image such as contrast, energy, and the use of fuzzy logic are considered by selecting the Harris threshold level.First, using each cover image, some images are obtained with different contrasts and a constant Harris threshold level; besides, in order to extract the above-mentioned features, the "Gray Level Co-occurrence Matrices" (GLCM) are calculated.Then, using the fuzzy logic, different security levels are presented for images.The images with different security levels could be stored in different image banks.Afterwards, to achieve lower detectability a stegangraphy against a steganalyzer technique, the difference were obtained between features of clean and stego images, and then to evaluate the performance of each feature, "Area Under the Curve" (AUC) values were calculated using the steganalyzer to the "Fisher Linear Discriminant" (FLD) classification.Actually, the cover bank is formed through cover classification based on an index pre-steganography security level.