基于大数据的圆对称扩频数字图像篡改盲检测  

Digital Image Tampering Blind Detection of Circular Symmetry Spread Spectrum Based on Big Data

作　　者：李兵[1] 赵明华[1] 王锋[1] LI Bing;ZHAO Ming-hua;WANG Feng(School of Computer Science and Engineering,Xian University of Technology,Xian Shanxi 710048,China)

机构地区：[1]西安理工大学计算机科学与工程学院,陕西西安710048

出　　处：《计算机仿真》2022年第4期419-422,427,共5页Computer Simulation

基　　金：西安理工大学教学研究项目(xjy1864)。

摘　　要：为了精准有效判别图像是否被恶意篡改,保证其图像的真实性,提出一种基于大数据的圆对称扩频数字图像篡改盲检测方法。在大数据环境下将伪随机序列当作水印,按照离散傅立叶转换特征融合通信扩频技术,将圆对称水印嵌入图像频域内;构建基于JPEG双重压缩效应的数字图像篡改模型,通过区间长度表达双重压缩前后系数转换的相关性,并将压缩处理后的图像存储为JPEG格式,正确定位被篡改位置;对图像块采取Radon变换及解析Fourier-Mellin变换,提取变换后的矩阵特性值,得到矩阵特性关联,深层次检测图像真实性。仿真结果表明,所提方法可大幅度提升篡改图像的检测精度及效率,鲁棒性强,拥有较高的实用性。In order to accurately and effectively judge whether the image has been maliciously tampered and ensure the authenticity of images, a method of digital image tampering blind detection based on big data for circular symmetry spread spectrum was proposed. In the big data environment, the pseudo-random sequence was used as the watermark, and then the discrete Fourier transformation feature fusion communication spread spectrum technology was used to embed the circular symmetry watermark into the image frequency domain. The digital image tampering model based on JPEG double compression effect was constructed. Moreover, the correlation of coefficient conversion before and after the double compression was expressed by the interval length. The compressed image was stored in JPEG format to locate the tampered position correctly. Radon transformation and analytic Fourier-Mellin transform were used to extract the characteristic value of the matrix, so that the feature correlation of the matrix was obtained. Finally, the image authenticity was detected in depth. Simulation results show that the proposed method can greatly improve the accuracy and efficiency to detect tampered images, with strong robustness and high practicability.

关 键 词：大数据 圆对称扩频数字图像 篡改盲检测 

分 类 号：TP367[自动化与计算机技术—计算机系统结构]