Improved data hiding capacity through repeated embedding using modified weighted matrix for color image

摘要

Till date, data hiding techniques using interpolated grayscale image were limited embedding capacity and visual quality. In this investigation, we have proposed a weighted matrix-based reversible data hiding scheme using the color image which provides image authentication and tampered detection. Here, we decomposed the original image into RGB color components and partitioned into (3 x 3) pixel blocks then we performed the sum of entry-wise multiplication operations using modified weighted matrix to embed the secret data repeatedly in each block. The data embedding locations are stored within an index file to enhance security, increase data hiding capacity, gain good visual quality, achieve reversibility, and confirm authenticity. The proposed data hiding scheme also improved both data embedding capacity up to 8.03 (bpp) and visual quality PSNR 50.03(dB). Finally, the scheme is compared with other existing methods and achieve a reasonably better performance in terms of visual quality and payload. Our scheme has been evaluated through various steganographic analyses and observed that the scheme is secure and robust against various attacks. The Proposed scheme successfully performed tampered detection and authentication but may partially recovered original cover image from tampered stego image.