基于视觉效应的自适应图像隐写算法

空间域中,本文根据光混合颜色立方体,通过计算欧式空间中的特征向量,提出了衡量主观不敏感区域的特征参数"色度"、"色偏差",根据设定的参数阈值筛选隐写视觉效果好的像素位置.算法通过分析原始像素值与待隐写数据的数字特征,自适应地将信息写到特征相近的像素点中,并记录下隐写位置表.通过理解隐写位置表,算法在不需要原始图像的情况下完成信息提取.同时,算法还提出了数据隐写协议,有效地避免了图像遭破坏后导致的信息误传.仿真结果表明,该隐写算法视觉效果好,利用空间域最常用的隐写分析算法分析法和RS分析法对图像进行隐写分析也无法准确判断图像是否含有隐写信息.

Adaptive Steganography Algorithm Based on Vision Effect

In the spatial domain,the characteristic parameters "chroma" and "colour difference" based on the theory of light mix are proposed to describe insensitive feeling through calculating the feature vector in Euclid space.The optimum location can be found by filtering the parameters.Using figure characters,we propose a novel algorithm to put information into image adaptively,so that it can maintain figure characters.Steganography information can be obtained without original image by recording information location.At the same time,the steganography protocol used in this algorithm avoids transmitting misinformation effectively.The theoretical analysis,confirmed by experimental results,shows that the proposed algorithm has a good vision effect.Neither χ2 analysis nor RS analysis can confirm whether the image has been disposed or not.