Optical image hiding based on chaotic fingerprint phase mask and diffractive imaging

This paper presents a novel optical image hiding scheme that employs chaotic fingerprint phase masks (CFPMs) and diffractive imaging. The proposed scheme encodes the secret image as a noise-like diffractive pattern using CFPMs positioned at various diffractive distances. Principal component analysis (PCA) is employed to obtain the score matrix of the noise-like diffractive pattern, and singular values are derived from the score matrix using singular value decomposition (SVD). Additionally, the host image is subjected to a 3-level discrete wavelet transform (3-LDWT) to extract low-frequency coefficient features. The same PCA and SVD operations are performed on these features and combined with the singular values of the secret image to complete the image hiding process. This scheme effectively balances the invisibility and robustness of the hiding scheme by embedding the low-dimensional features of the secret image into the low-frequency features of the host image. Moreover, the inclusion of biometric features enhances the association with the key and authorized users, while the incorporation of chaos mapping and diffractive parameters provides additional security for the proposed scheme. The feasibility, efficiency, security, and robustness of the scheme are verified through a series of rigorous experiments.