Study On Information Hiding Techniques Based On Human Visual System

HVS-based BPCS and PVD steganography technologies have the characteristics of large embedding capacity and good imperceptibility, so they play an important role in the field of information hiding. But when the steganograpgy technologies have not net grown up, the related steganalysis technologies have been gradually proposed and threatened the security of the traditional steganography technologies. On account of the shortcomings of the BPCS and PVD steganography, the paper proposed two improved steganography algorithms and the algorithms can effectively resist to the detection of the corresponding steganalysis. The information hiding algorithm based on IWT was remarkable for good imperceptibility and strong anti-attack capability, but limited embedding capacity. On the basis of the previous studies and the characteristics of HVS, the embedding capacity of information algorithm based on IWT bas been significantly enhanced. Furthermore, the imperceptibility of algorithm bas also been improved.The first, to enhance the ability of anti-statistical analysis of BPCS steganograpgy, the paper proposed a dynamic threshold BPCS steganography algorithm based on image classification. The algorithm utilized setting dynamically the bit-plane complexity threshold and the similarity of bit-plane block of carrier image and secret information to improve the imperceptibility of the BPCS steganograpgy algorithm. To decrease the intensity of complexity histogram, the complexity threshold of bit-plane blocks was determined by the basic bit-plane complexity threshold and random number, so the algorithm could effectively resist to the BPCS steganalysis.The second, to enhance the imperceptibility of information hiding algorithm based on IWT, the paper proposed an IWT color information hiding algorithm based on HVS. The algorithm solved the issue of gray overflow because of information hiding and determined the classified result and embedding capacity of wavelet coefficient based on the relation of image blocks and wavelet coefficient blocks. To enhance the imperceptibility, the algorithm introduced the concept of BPCS steganography complexity into it. The third, to enhance the ability of anti-statistical analysis of PVD steganograpgy, the paper proposed a hybrid algorithm of PVD and LSB based on texture analysis. The algorithm generated random integers to adjust the boundary of difference value intervals for pixel pairs of PVD steganography part and calculated optimum difference values after steganography. To enhance the quality of the stego image, the algorithm used the Optimal Pixel Adjustment Process to adjust the gray-value of pixel blocks after LSB steganography. The algorithm not only improved imperceptibility and capacity, but also could resist to the PVD steganalysis.