图像隐藏信息与图像噪声分辨研究

针对图像经过信息隐藏与经过噪声干扰具有不同的特性，从图像隐藏信息和加噪的数学原理入手，引入图像直方图和差分直方图的模型，在图像小波分解高频系数差分直方图的数学模型基础上，利用高频系数差分数值的方差对图像是否隐藏信息或被噪声干扰进行分辨，给出了具体的计算判断方法。最后，在标准图像集上进行实验。结果表明，即使在图像直观统计参数基本相同的情况下，该方法也能明显分辨出两种不同的干扰现象，取得良好结果。

Distinguishing image data hiding and image noise

Since the information embedded image has different features with the noised picture, the principle of the information hiding and noise jamming image is analyzed and the image histogram model and the difference histogram model are introduced. Based on the model of the high frequency coefficients difference histogram of the image wavelet decomposition, the variance of the high frequency coefficients difference value is used to distinguish the image data between hiding and noise jamming, and the concrete computing judgment method is given. The experiment on the standard images shows that this method can distinguish the noise image and embedded image very well, even if their intuitionistic statistics are approximately equal.