Authors: Guangjie Liu; Weiwei Liu
Addresses: School of Automation, Nanjing University of Science and Technology, Nanjing, 210094, China. ' School of Automation, Nanjing University of Science and Technology, Nanjing, 210094, China
Abstract: In this paper, an adaptive matrix embedding method based on LSB matching is proposed for greyscale images. The adaption mechanism of the method is used to embed less message bits in regions with high local correlation and embed more in regions with low local correlation. The 2 × 4 pixel block is taken as a cover unit. The seven pixels of its neighbour left column and up row are utilised to estimate the local correlation of the block. The estimated local correlation is further used to determine the message bit amount and choose the corresponding matrix embedding strategy. And LSB matching is performed to change the parity of the pixel needing modification determined by the matrix embedding. The experimental results indicate that in the test image database the new method can improve the PSNR value as high as about 3.75 dB as compared with simple LSB matching. And our method also has stronger resistance ability against Ker's adjacency HCF-COM detector than simple LSB matching.
Keywords: matrix embedding; LSB matching; LSBM; adaptive steganography; grey-scale images.
International Journal of Multimedia Intelligence and Security, 2011 Vol.2 No.3/4, pp.238 - 251
Received: 08 May 2021
Accepted: 12 May 2021
Published online: 06 Jan 2012 *