Artificial Intelligence in Digital Holographic Imaging: Technical Basis and Biomedical Applications

ISBN-13: 9780470647509 / Angielski / Twarda / 2022

Moon, Inkyu

Opis

This book presents a ground–breaking intelligent system for fast and non–invasive microbial identification using 3D optical imaging methods and high throughput algorithms for automatic analysis of 3D and 4D microscopic image data, as well as analysis of microscopic imaging towards a basic understanding of biological specimens.

Spis treści

Part I. Digital Holographic Microscopy (DHM)1. IntroductionReferences2. Coherent optical imaging2.1 Monochromatic fields and irradiance2.2 Analytic expression for Fresnel diffraction2.3 Transmittance function of lens2.4 Geometrical imaging concepts2.5 Coherent imaging theoryReferences3. Lateral and depth resolutions3.1 Lateral resolution3.2 Depth (or axial) resolutionReferences4. Phase unwrapping4.1 Branch cuts4.2 Quality-guided path-following algorithmsReferences5. Off-axis digital holographic microscopy5.1 Off-axisdigital holographic microscopy designs5.2 Digital hologram reconstructionReferences6. Gabor digital holographic microscopy6.1 Introduction6.2 MethodologyReferencesPart II. Deep Learning in DHM Systems7. IntroductionReferences8. No-search focus prediction in DHM with deep learning8.1 Introduction8.2 Materials and methods8.3 Experimental results8.4 ConclusionsReferences9. Automated phase unwrapping in DHM with deep learning9.1 Introduction9.2 Deep learning model9.3 Unwrapping with deep learning model9.4 ConclusionsReferences10. Noise-free phase imaging in Gabor DHM with deep learning10.1 Introduction10.2 A deep learning model for Gabor DHM10.3 Experimental results10.4 Discussion10.5 ConclusionsReferencesPart III. Intelligent DHM for Biomedical Applications11. IntroductionReferences12. Red blood cells phase image segmentation12.1 Introduction12.2 Marker-controlled watershed algorithm12.3 Segmentation based on marker-controlled watershed algorithm12.4 Experimental results12.5 Performance evaluation12.6 ConclusionsReferences13. Red blood cells phase image segmentation with deep learning13.1 Introduction13.2 Fully convolutional neural networks13.3 Red blood cells phase image segmentation via deep learning13.4 Experimental results13.5 ConclusionsReferences14. Automated phenotypic classification of red blood cells14.1 Introduction14.2 Feature extraction14.3 Pattern recognition neural network14.4 Experimental results and discussion14.5 ConclusionsReferences15. Automated analysis of red blood cell storage lesions15.1 Introduction15.2 Quantitative analysis of red blood cell 3D morphological changes15.3 Experimental results and discussion15.4 ConclusionsReferences16. Automated red blood cells classification with deep learning16.1 Introduction16.2 Proposed deep learning model16.3 Experimental results16.4 ConclusionsReferences17. High-throughput label-free cell counting with deep neural networks17.1 Introduction17.2 Materials and methods17.3 Experimental results17.4 ConclusionsReferences18. Automated tracking of temporal displacements of red blood cells18.1 Introduction18.2 Mean-shift tracking algorithm18.3 Kalman filter18.4 Procedure for single RBC tracking18.5 Experimental results18.6 ConclusionsReferences19. Automated quantitative analysis of red blood cells dynamics19.1 Introduction19.2 Red blood cell parameters19.3 Quantitative analysis of red blood cell fluctuations19.4 ConclusionsReferences20. Quantitative analysis of red blood cells during temperature elevation20.1 Introduction20.2 Red blood cell sample preparations20.3 Experimental results20.4 ConclusionsReferences21. Automated measurement of cardiomyocytes dynamics with DHM21.1 Introduction21.2 Cell culture and imaging21.3 Automated analysis of cardiomyocytes dynamics21.4 ConclusionsReferences22. Automated analysis of cardiomyocytes with deep learning22.1 Introduction22.2 Region of interest identification with dynamic beating activity analysis22.3 Deep neural network for cardiomyocytes image segmentation22.4 Experimental results22.5 ConclusionsReferences23. Automatic quantification of drug-treated cardiomyocytes with DHM23.1 Introduction23.2 Materials and methods23.3 Experimental results and discussion23.4 ConclusionsReferences24. Analysis of cardiomyocytes with holographic image-based tracking24.1 Introduction24.2 Materials and methods24.3 Experimental results and discussion24.4 ConclusionsReferences25. Conclusion and future work

O autorze

Inkyu Moon, PhD, is Professor in the Department of Robotics and Mechatronics Engineering at Daegu Gyeongbuk Institute of Science & Technology (DGIST), South Korea.