Steganalysis for LSB inserts in low stego-payload artificial color images

Abstract

In this study a new method of steganalysis is presented for detecting and locating a hidden message embedded using LSB replacement as a steganography algorithm. The method is applied for analysing artificial color images of a gradient fill and shows high detection and locating accuracy in low stego-payload of 25 to 10 percentage. The method is based on analysing pixel combinations of an LSB map to classify unique and non-unique pixel combinations and finding the largest rectangle algorithm for, respectively, detecting and locating purposes. To exclude noisy data from further consideration, a pre-processing filter is developed and presented.