中文摘要：

为提高图像隐写的统计安全性,根据图像内容的最小化嵌入失真原则,在空域通用小波相对失真方法 （spatialuniversal wavelet relative distortion,S-UNIWARD）的基础上,提出一种结合边缘检测和校验格编码（syndrome trellis code,STC）的图像自适应空域隐写术。利用基于Canny算子的边缘检测得到图像的边缘区域,根据S-UNIWARD定义图像像素的嵌入失真,通过实验设置失真阈值来选择图像的纹理区域,使用STC在边缘区域和纹理区域对秘密信息进行嵌入。实验结果表明,该算法提高了S-UNIWARD在图像边缘区域和纹理区域的嵌入精度,提高了算法安全性。

英文摘要：

To improve the statistical security of image steganography,according to the principle of minimizing suitable defined distortion in the content of digital image,an adaptive spatial steganographic algorithm was proposed on the basis of spatial-universal wavelet relative distortion（S-UNIWARD）,in combination with edge detection and syndrome trellis codes（STC）.Canny＇s edge detection was used to select the edge region of the cover image.The embedding distortions of the edge pixels were defined according to S-UNIWARD,and the image＇s texture region was selected by setting the distortion threshold through experiments.The secret message was embedded in the pixels where were edge and texture regions using STC.Experimental results illustrate that the proposed method improves not only the accuracy of embedding in edge and texture regions,but also the security of the steganography.