基于DCT域的数字图像隐写容量归一化方法  

Digital Image Steganography Capacity Normalization Method Based on DCT Domain

作　　者：吴煌 李凯勇[2] WU Huang;LI Kai-yong(School of Electronic Information and Computer Engineering,Sichuan Institute of Industrial Technology,Deyang Sichuan 618500,China;Qinghai Nationalities University Physics and Electronic Information Engineering,Xining Qinghai 810007,China)

机构地区：[1]四川工业科技学院电子信息与计算机工程学院,四川德阳618500 [2]青海民族大学物理与电子信息工程学院,青海西宁810007

出　　处：《计算机仿真》2021年第8期207-211,共5页Computer Simulation

摘　　要：针对现有方法存在的隐写容量较小问题,提出基于DCT域的数字图像隐写容量归一化方法。结合隐写术不可感知性、鲁棒性等特点,根据对"囚犯问题"的分析,从信息嵌入、载体传输与信息提取三方面构建隐写系统模型;对载体图像做DCT变换,使其呈现频谱均匀特征;利用序列密码方式对秘密信息加密,产生混沌序列,确定嵌入的最佳位置、时机与强度等条件,将隐秘信息嵌入到量化系数中,获得秘密图像;分别提取DCT块内与块间容量特征,通过改变像素中1位容量,利用突显缩放方法确保容量特征子集具有相同1-范数,实现容量归一化。仿真结果表明,上述方法能够有效提高隐写容量,获取的秘密图像信噪比更高。This paper proposes a normalization method of digital image steganographic capacity based on the DCT domain for improving the steganographic capacity of traditional methods.First of all,based on the analysis of the "prisoner problem" and the imperceptibility and robustness of steganography,the steganography system model was established from information embedding,carrier transmission,and information extraction.Secondly,in order to show the uniformity of the spectrum,the carrier image was transformed by DCT.Then,the sequence cipher was used to encrypt the secret information,and the chaotic sequence was generated to determine the best location,time,and strength of embedding.The secret information was embedded into the quantization coefficient to obtain the secret image.The capacity features within and between DCT blocks were extracted,respectively.Eventually,changing the 1-bit capacity of pixels,and using highlight scaling to ensure that the capacity feature subset had the same 1-norm,the capacity normalization was achieved.Simulation results show that this method can improve steganography capacity and obtain secret images with a high signal-to-noise ratio.

关 键 词：离散余弦变换 数字图像 隐写容量 归一化处理 突显缩放 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]