| 1. | We used this system to check some representative images on the web , we could got a conclusion that most encode images is easy to decode . without right theory they can ’ t build a robust encoded image . and through thorough experiments by this platform i have concluded such points : single alpha encoded is useless , we must encode all characters as whole ; confusion points are useless , we must use confusion lines which look like stroke ; enlarge base of character such as chinese character , not only digital number and alpha
并且通过该平台的广泛实验总结出增强图片加密强度的几个要点:只加强单个字的加密是不行的,必须把整体作为加密对象,增强字与字之间的干扰;单纯的噪点作用不大,具有笔画特征的干扰线作用更大;应该引入一些随机变量,比如每副图片的字个数随机,加密方法随机选择;增大可选词汇的基数,可以考虑使用汉字,而不单纯是数字和英文。 |
