一种基于FCM和HVS的隐写方法

Steganographic Method Based on FCM and HVS

为了提供较大的秘密信息嵌入量并保持良好的载密图像质量,根据对比度和梯度敏感度,并利用模糊聚类的分类能力,提出了一种基于FCM(模糊聚类)和HVS(人眼视觉特性)的图像隐写方法。该方法将对比度和梯度敏感度作为特征向量,并通过模糊C均值聚类将像素分为视觉敏感类与视觉不敏感类,采用模算子隐写将较多秘密信息嵌入属于视觉不敏感类的像素,而将较少秘密信息嵌入属于视觉敏感类的像素。实验结果表明,与SMFH算法相比,该算法有更大的嵌入量。

To provide larger capacity of the hidden secret data and to maintain a good visual quality of stego-image,in terms of contrast and gradient sensitivity,the paper puts forward a steganographic method based on FCM (Fuzzy C-means Clustering) and HVS (Human Visual System),by making use of the classification ability of fuzzy clustering. This method uses contrast and gradient sensitivity as feature vector,and divides pixels into the sensitive class and insensitive class by FCM. More secret data can be embedded in the pixels of insensitive class with steganography based on modulus,and data with little secret can be embedded in the pixels of sensitive class. The experimental results show that the proposed method hides much more information compared with SMFH.