Towards improving the security of image steganography via minimizing the spatial embedding impact

With the continuous advances in computer technology and electronics, image steganography, an important branch of information hiding aimed at hiding secret information in innocent digital images for covert communication, has also been developed significantly in recent years. Recent studies on image steganography have shown that non-additive distortion steganography schemes, e.g., Synch, CMD and Dejoin-J, can improve the security of additive steganographic distortion functions. These schemes are basically designed based on some heuristic experiences, whose essence can be attributed to implicitly reducing the spatial embedding impact. In view this, we develop a general Intra-block Modification Optimization (IbMO) strategy to improve the security of both spatial and JPEG image steganography via directly minimizing the spatial embedding impact, wherein the metric for spatial embedding impact is explicitly defined as the summation of absolute difference of noise residuals between cover and stego image. Note that, to apply it to JPEG image steganography, the IbMO strategy will be integrated into a newly proposed BBC-based Intra-block Joint Modification (IbJM) scheme with five-element crossing neighborhood. Experimental results indicate that the proposed IbMO strategy can obviously improve the security performances of additive distortion steganography schemes. Specifically, the IbMO strategy significantly outperforms Synch and rivals CMD on HiLL and MiPOD. The newly proposed IbJM scheme also shows better overall performance with less computational overhead than DeJoin-J on JUNIWARD, GUED, and BET, and its security performance can be further improved after integration with IbMO strategy.(c) 2022 Elsevier Inc. All rights reserved.