Artificial Intelligence in Digital Holographic Imaging

11Introduction

DHM is a new and highly promising approach to identify cellular biomarkers, particularly when it is combined with AI or deep‐learning techniques for biomedical applications [1–10]. DHM is also a contact‐less and label‐free method. Thus, a sample can be studied without damaging it. Part III provides an overview of some of the recently published work on AI or deep‐learning techniques in holographic image analysis as a tool to study the intracellular content and morphology of live cells [1,11–30]. The overview focuses on the automated phenotypic analysis of live RBCs and cardiac cells via intelligent DHM.

Chapters 12 through 20 introduce an automated phenotyping platform based on DHM for the quantitative analysis of RBCs. Chapters 12 and 13 introduce RBC phase image‐segmentation techniques essential for automated RBC analysis. Chapters 14–16 demonstrate that integrating DHM techniques integrated AI enable scientists to obtain rich, quantitative information about the structure of RBCs in non‐invasive, real‐time conditions for automatic phenotypic classification of RBCs. Chapter 17 shows that deep‐learning DHM can also rapidly detect and count multiple cells in hologram images at the single‐cell level, which is needed for high‐throughput cell counting. Chapter 18 introduces a tracking algorithm to locate a single RBC through RBC image sequences obtained with time‐lapse DHM and dynamically monitor its biophysical cell parameters. Chapter 19 introduces methods to quantitatively ...