基于混沌映射的图像认证技术

针对脆弱水印认证算法的篡改定位精度及安全性问题,提出了一种篡改检测和定位分离的图像认证方案.将原始图像的像素灰度值映射为混沌初值,经过若干次迭代生成检测水印和定位水印,然后将其嵌入到像素灰度值的低两位.利用混沌对初值的极端敏感性,能够精确地定位对含水印图像的篡改,并且水印提取无需宿主图像.实验结果表明本算法能够检测出任何对图像像素值的改变和对图像完整性的破坏,同时可将篡改定位到单个像素,有效地实现了对数字图像的精确认证.

Image Authentication Technique Based on Chaotic Map

Being aimed at the tamper location accuracy and the security problem of existing fragile watermarking authentication algorithms, an image authentication scheme is presented which can separate tamper detection and tamper location. Let the grayscale value of the host image as the initial value of the chaotic mapping to generate detection watermarking image and location watermarking image, then embedded them into the Hypo-LSB and LSB of each pixel. The proposed scheme can locate the tampers on the watermarked image accurately by using the high sensitivity on initial value of chaotic mapping. Moreover, no host image is needed in watermark extraction. The experiment results demonstrate that the fragile watermarking arithmetic can detect any change to image pixel value and any destroy to image completeness, and also can locate tampered area in pixel-level accurately.