Deep Learning-based Algorithms for Nucleus Segmentation: An Overview

Nucleus segmentation is a crucial task in biological and medical imaging, facilitating in-depth investigations into cellular nuclear morphology, cell tracking, and cellular phenotype classification. With the advent of deep learning, we've seen significant advancements in this domain, leading to higher accuracy and the ability to process images at an unprecedented scale. In this overview, we specifically discuss the application of various deep learning methods in nucleus segmentation and present various evaluation metrics. Nevertheless, we also discuss the current challenges being faced and try to outline the future directions of development.