The Research Of A Image Steganography Based On Human Visual System

Steganography is the means of concealed communication and it takes an important role in military affairs, information and security of country. The basic demand for steganography is providing good security and large embedding capacity. The steganography based on HVS has good visual imperceptibility, and its payload is the most in all steganographic methods. We try to apply wavelets, Self Organizing Feature Maps and Fuzzy clustering to image steganography, based on Human Visual System, proposing two steganographic methods with better performance.First, in view of contrast and gradient sensitivity, and exploiting the classified ability of fuzzy clustering, a novel steganographic method based on FCM (Fuzzy C-means Clustering) and HVS (Human Visual System) is presented. The contrast and gradient sensitivity are used as feature vector, and by the way of FCM, the pixels are classified into the sensitive class and the less sensitive class. More secret data can be embedded in the pixels of less sensitive class with steganography based on modulus, and less secret data can be embedded in the pixels of sensitive class. From the experimental results, compared with SMFH, the proposed method hides much more information.Second, in view of contrast and texture sensitivity, and exploiting the classified ability of self-organizing competitive NNS, a novel stegano-graphic method on wavelet domain based on SOM (Self-organizing Feature Map) and HVS (Human Visual System) is presented. Firstly, an image is divided into blocks, and the sum of wavelet contrast and the variance of wavelet coefficients are features. Then, those blocks are classified into 3 classes by SOM. Finally, secret information is embedded into the wavelet coefficients with steganography based on modulus. From the experimental results, compared with WHS, the proposed method hides much more information and maintains a good visual quality of stego-image.