Detection Methods for LSB Embedding Based on Laplacian Statistics

Two new steganalysis methods for the Least Significant Bit (LSB) embedding technique are proposed based on the analysis of the laplacian statistics of image pixel. The relation between the current pixel and its four-neighborhood pixel average is defined as statistics, then we propose the method 1, which can detect the existence of hidden messages, and the method 2, which can accurately estimate the amount of hidden messages. These two methods are based on the analysis of effects brought by message embedding, embedding twice and LSB plane flipping. These two methods have remarkable physical significances and can be implemented conveniently. Experimental results show that the estimating precision of method 2 is better than that of the RS method if the embedding ratio is not less than 20%.