Artificial intelligence for pathology

Abstract

Advances in artificial intelligence (AI), especially in deep learning, improve pathological image analysis in basic, translational, and clinical research and in routine clinical practice. Deep learning is currently the dominant technique among the best solutions for many tasks in digital pathology. This chapter provides a general overview of different applications of deep learning in pathological image analysis, such as image classification, object detection, image segmentation, stain normalization, and image superresolution. It summarizes deep learning achievements and identifies the contributions in specific tasks. In addition, it discusses open challenges and potential directions of deep learning-based pathological image computing and presents barriers to clinical adoption of AI in digital pathology.