Simulation of biomedical signals and images using Monte Carlo methods for training of deep learning networks
Deep Learning Techniques for Biomedical and Health Informatics
Electron Science Research Institute
© 2020 Elsevier Inc. All rights reserved. High accuracy supervised deep learning methods require massive data with accurate ground truth. However, in biomedical applications, exact measurement of the ground truth is often impractical or even impossible. An important avenue to generate data with ground truth is simulation of the biomedical or imaging process. In this chapter, Monte Carlo methods are proposed as a useful set of tools to generate physics-based simulated signals and images.