Document Type : Regular article

Authors

• Ramya N ¹
• Hossein Rashmanlou ²
• A. A. Talebi ³
• Farshid Mofidnakhaei ⁴
• Mostafa NouriJouybari ⁵

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

² Canadian Quantum Research Center, 106-460 Doyle Ave, Kelowna, British Columbia V1Y 0C2, Canada;

³ Department of Mathematics, University of Mazandaran, Babolsar, Iran;

⁴ Department of Physics, Sari Branch, Islamic Azad University, Sari, Iran;

⁵ Payame Noor University

Abstract

Digital holography has emerged as a powerful label-free imaging modality for quantitative interrogation of microscale flow phenomena in microfluidic and biomedical systems; however, conventional reconstruction methods remain limited by ill-posed inverse formulations, noise sensitivity, twin-image artifacts, and substantial computational cost. Recent advances in artificial neural networks (ANNs) have transformed this landscape through data-driven, physics-informed, and hybrid reconstruction strategies that improve phase retrieval, denoising, particle tracking, and velocity estimation while enabling near real-time diagnostics. Following a PRISMA-guided critical synthesis, this review systematically examines the evolution of ANN-assisted digital holography, spanning optical foundations, conventional computational reconstruction, deep learning frameworks, and emerging physics-constrained models. Existing approaches are organized into physics-based, data-driven, and hybrid paradigms, and critically bench marked in terms of reconstruction fidelity, computational efficiency, automation readiness, and diagnostic applicability. Particular emphasis is placed on applications in flow cytometry, microcirculation analysis, label-free disease detection, and intelligent lab-on-a-chip diagnostics. The review further identifies persistent challenges involving model generalization, annotated data scarcity, explainability, and multiphysics integration, while highlighting emerging opportunities in physics-informed learning, autonomous holographic diagnostics, and digital twin-enabled flow monitoring. By consolidating current advances and outlining a strategic research roadmap, this work establishes a unified framework for next-generation intelligent quantitative flow diagnostics through the convergence of digital holography and artificial intelligence.

Keywords

• Digital Holography
• Quantitative Phase Imaging
• Microfluidics
• Biomedical Flow Imaging
• Artificial Neural Networks
• Physics-Informed Neural Networks
• Lab-on-a-Chip

Main Subjects

• Physics