Abstract

Data hiding is a branch of information security that is popularly used for digital watermarking and steganography. Traditional data hiding methods are not suitable for some information sensitive domains such as medical imaging, biometric data protection, military etc. because they can permanently damage the cover image. These fields do not tolerate the cover image to be modified even slightly. There is a need of reversible data hiding in such domains that allows to completely recover the cover-image after the retrieval of secret message. Pixel Value Ordering (PVO) for hiding data reversibly has gained much popularity because it maintains high quality of the resultant stego image. Improved Pixel Value Ordering (IPVO), an extension of PVO, is a high-fidelity method that uses sorting and Prediction Error Expansion(PEE) for hiding data. In IPVO, the cover-image is segregated to form segments of fixed size. These segments, called blocks, are sorted in increasing order of pixel intensities. Then, the prediction error for the largest and the smallest pixels are computed using the second-maximal and the second-minimal pixels, respectively. Thus, the largest-valued and the minimumvalued pixels are utilized to embed data. Each block can embed at most two bits. The data carrying capacity is constrained by the number of blocks. In IPVO, the blocks that are already sorted may produce prediction error of -1 that is not utilized for data embedding. IPVO only makes use of prediction errors 0 and 1, and the error -1 is ignored. In this paper, an improvement of the existing IPVO method is proposed. The proposed method makes use of a novel block partitioning scheme in such a way that a block consists of three pixels. Moreover, the proposed method utilizes a modified IPVO strategy to utilize the prediction error of -1 to hide the data. Hence, the proposed method achieves a high data carrying capacity when compared to the other unconventional methods that use PVO.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call