Conclusion and Future Work

Abstract

This conclusion presents some closing thoughts on the key concepts discussed in the preceding chapters of the book. This book discusses applications of deep learning in digital holographic cell imaging and digital holographic microscopy (DHM)-based phenotypic analysis methods. It presents a deep-learning convolutional neural network with a regression layer as the top layer to estimate the best reconstruction distance. The book describes a new deep-learning model that can automatically reconstruct unwrapped, focused phase images by combining digital holography and a generative adversarial network for image-to-image translation. It demonstrates the potential of novel approaches to study live red blood cells (RBCs) by integrating DHM with deep learning, which achieved good segmentation and classification accuracy with a Dice coefficient of 0.94 and a high-throughput rate of about 152 cells per second. Moreover, the holographic image-based deep-learning models could be applied to identifying morphological changes that occur in RBCs during storage.