Document Type : Regular article

Authors

• Ramya N ¹
• Janusz Kacprzyk ²
• Hossein Rashmanlou ³
• Afshin Fayyaz Movaghar ⁴
• Farshid Mofidnakhaei ⁵

¹ Department of Mathematics, Kongu Engineering College, Erode, India

² Intelligent Systems Laboratory, Systems Research Institute, Poland

³ Canadian Quantum Research Center, 106--460 Doyle Ave, Kelowna, British Columbia V1Y 0C2, Canada

⁴ Department of Statistics, University of Mazandaran, Babolsar, Iran

⁵ Department of Physics, Sari Branch, Islamic Azad University, Sari, Iran

Abstract

Digital holography has evolved from a conventional wavefront-recording technique into an intelligent computational imaging paradigm capable of recovering complex optical fields and extracting quantitative information from phase-sensitive measurements. The rapid integration of artificial intelligence (AI), fractional Fourier transform (FrFT) methods, phase-retrieval algorithms, quantitative phase imaging (QPI), and digital holographic microscopy (DHM) has substantially expanded the capabilities of holographic systems in biomedical imaging, optical metrology, and label-free diagnostics. This review provides a comprehensive analysis of the recent development of AI-enhanced digital holography, with particular emphasis on the mathematical foundations of hologram formation, Fourier-domain and Fresnel reconstruction, fractional-domain processing, iterative phase retrieval, deep learning architectures, and quantitative phase analysis. Classical and emerging computational strategies are critically compared in terms of reconstruction fidelity, computational efficiency, phase accuracy, interpretability, and application-specific adaptability. A numerical digital holography framework is further considered to examine the complete processing chain from complex object generation and off-axis hologram formation to FFT-, Fresnel-, and FrFT-based reconstruction, phase retrieval, optical path difference estimation, and thickness mapping. The comparative analysis demonstrates that FFT-based reconstruction provides the strongest object-structure preservation under the investigated conditions, whereas Fresnel propagation offers greater physical relevance for depth-dependent reconstruction and numerical refocusing. The FrFT framework provides a flexible intermediate representation but remains highly dependent on fractional-order optimization. In parallel, AI-based approaches demonstrate considerable potential for automated phase recovery, aberration correction, image enhancement, object classification, and quantitative cellular analysis. The review identifies current limitations associated with data dependence, model generalization, phase ambiguities, computational complexity, experimental calibration, and the lack of standardized benchmarks. Finally, future research directions are outlined toward physics-informed learning, self-supervised holographic reconstruction, multimodal quantitative imaging, adaptive fractional-domain processing, and explainable AI-enabled digital holographic systems for reliable biomedical and industrial applications.

Keywords

• Digital Holography
• Artificial Intelligence
• Fractional Fourier Transform
• Quantitative Phase Imaging
• Digital Holographic Microscopy
• Deep Learning

Main Subjects

• Physics