【摘要】 Arnold变换、Baker变换以及幻方变换等传统的图像置乱方法一般应用于正方形图像的处理,而在采用传统方法对宽高不等的普通矩形图像进行置乱时,由于处理对象普遍不具备变换周期,故难以直接运用上述方法。文章采用基于区域一致性的图像插值技术对图像进行预处理,然后对图像矩阵进行置乱与快速的逆置乱,并通过一系列客观的评价方法对图像处理的过程及结果进行分析,实验结果表明文中所提方法是可行有效的。更多还原

【Abstract】 Traditional image scrambling methods such as Arnold transformation,Baker transformation and the Magic Quadrant transformation are usually applied to the processing of equilateral images.But for the non-equilateral ones,the methods mentioned above could be hardly utilized directly because the transformation periods do not commonly exist in these ones during the transformation process.The image interpolation is adopted to do pre-process in order to make traditional scrambling methods be quickly and directly applied to non-isometric image scrambling.A series of objective evaluation methods is employed to analyze the process and quality of image processing.The experimental results show the feasibility and validity of the raised method.更多还原