Abstract: Image scrambling is a common method in information hiding with purpose of improving robustness and imperceptibility of image by scrambling image’s information order.How to evaluate the image scrambling degree is the important issue of research.This paper describes a new evaluation method of scrambling degree based on the description of image quality which measures the image features of texture primitive scale and continuous probability of gray level.Experimental results show that the proposed method can well evaluate the scrambling degree and is consistent with human subjective vision.

摘要： 图像置乱是信息隐藏中最常用的一项技术，其目的是通过将图像的信息次序打乱来提高隐蔽载体的鲁棒性和隐蔽性。如何评价图像置乱的程度，是信息隐藏研究的重点问题之一。利用图像纹理基元行程特征和灰度级的连续概率特征对图像质量的描述方法，提出了基于图像纹理特征的图像置乱程度衡量方法，实验表明，该方法能较好地评价图像的置乱程度，与人的主观视觉评价有较好的一致性。