In this paper, a novel biometric image encryption scheme is proposed for securing iris templates through a hybrid integration of phase-only computer-generated holography (CGH) in the Fresnel domain, a chaotic umbrella map, unequal modulus decomposition (UMD) and a two-dimensional non-separable linear canonical transform. The phase-only computer-generated holography encodes the input iris data into a complex hologram. The chaotic umbrella map generates highly sensitive keys and significantly expands the key space. Whereas unequal modulus decomposition enhances confusion by diffusing input components. The two-dimensional non-separable linear canonical transform provides multi-parameter tunability and a strong non-linearity. Together, these transformations ensure a large key space, of making brute-force attacks infeasible. Performance analysis shows the scheme’s effectiveness, with entropy value of 7.9964 bits, near-zero correlation between encrypted and original images , and PSNR around 90 dB for decrypted images, ensuring near-lossless recovery. The encryption process is computationally efficient, averaging 1.12 seconds per encryption cycle. Additionally, the scheme exhibits strong resilience to differential attacks, with an average Number of Pixels Change Rate of 99.6539% and Unified Average Changing Intensity of 33.2106%.The proposed scheme is also resilient to noise, cryptographic attacks, and brute-force attempts, making it well-suited for real-world biometric security. By integrating advanced transformations with a strategically optimized encryption workflow, this work establishes a significant advancement over existing optical and digital encryption frameworks, providing an unprecedented combination of security, efficiency, and high-fidelity image reconstruction.
Rakheja, P., & Kaur, A. (2025). Robust phase-only hologram encryption for Iris biometrics using chaotic umbrella mapping and 2D NSLCT transform. Optics Communications, 132192.
