Advance Steganography Using Dynamic Octa Pixel Value Differencing

The practice of concealing messages or information within redundant medium is called as steganography. The various types of steganography depends on the choice of redundant medium, which are text, image, audio and video. In image steganography the secret message are embedded into the pixels of cover image to produce stego image i.e. image embedded with secret message bits. Least Significant Bit technique is one of the most widely used image steganography technique but it has many drawbacks like image distortion, detectability, susceptibility to cryptanalysis. To overcome these drawbacks Pixel Value Differencing technique was introduced. In PVD three neighboring pixels are compared with target pixel and there difference is calculated which in turn decides how many bits to embed in the pixel pair with the help of static range table. It minimizes the distortion and increase the storage capacity. But comparing three neighboring pixels is not enough as other surrounding pixels are not taken into consideration for data embedding. The range table is also static consisting of predefined sets of ranges. The proposed system compares all eight neighboring pixels to enhance data embedding capacity. Also, it uses a dynamic range table values of which changes as per the resolution of the image which helps to reduce the distortion. System also makes use of LSB + PVD approach to increase the embedding capacity, where LSB is used in smooth areas of image and PVD is used in edgy areas. Cryptographic measures are also added into system to provide additional data security and data integrity.