Optical fragile zero-watermarking scheme based on chaotic facial phase masks, single-shot digital holography, and...

Since most of the current optical image watermarking schemes generally embed the encoded watermark into the host image, they are not suitable for applications that require high integrity of the host image. Additionally, most of the existing optical image watermarking schemes are designed for robust watermarking and are unable to detect whether the image has been tampered with. Furthermore, in current optical image watermarking schemes based on human biometrics, the biometric features used in the encoding and decoding process of the watermark must be completely identical; otherwise, the useful information of the original watermark cannot be decoded. To address these issues, this paper proposes an optical fragile zero-watermarking scheme based on chaotic facial phase masks, single-shot digital holography, and canonical sine transform. In this proposed scheme, the chaotic facial phase masks are generated by a human face and a chaotic system. The original watermark is encoded into a noise-like image by chaotic facial phase masks and single-shot digital holography. Then, the encoded watermark is XORed with the canonical sine transform features of the host image to generate the zero-watermark. This scheme is suitable for applications that require high integrity of the host image. The proposed optical fragile zero-watermarking scheme does not require any modification to the original image, thus ensuring the integrity of the original image while effectively detecting whether it has been tampered with. Moreover, the key of this scheme is associated with the user's identity, so the proposed fragile watermarking scheme has extremely high security. Additionally, this scheme does not require the storage of facial biometric templates. Even if there are slight deviations in the user's facial features during encoding/decoding, the original watermark's valid information can still be successfully decrypted. Experimental results demonstrate that the proposed fragile zero-watermarking scheme exhibits strong fragility against common image processing attacks, geometric attacks, and tampering attacks.