Chapter Fifteen - Deep Learning

Abstract

Recent advances in high-performance computing enable training of highly complex neural networks for image segmentation and analysis. Deep neural networks are an emerging technique demonstrating unprecedented performance in a number of microscopy tasks. In particular, convolutional neural networks are suitable for learning underlying spatial features in images, and using this information for segmentation, classification, and improved resolution. In this chapter, we review trending topics in deep neural networks and summarize useful tips for network deployment for microscope image analysis. Finally, we discuss current challenges and potential trends in microscope image analysis using deep learning.