Published in:
07-12-2021 | Editorial
Deep learning-based attenuation map generation and correction; could it be useful clinically?
Authors:
Ananya Singh, MSc, Robert J. H. Miller, MD
Published in:
Journal of Nuclear Cardiology
|
Issue 6/2022
Login to get access
Excerpt
Deep learning (DL) is a branch of machine learning characterized by a multi-layered learning approach. Convolutional neural networks (CNNs) are an example of DL which are commonly applied to imaging data.
1,2 DL models are particularly well suited to direct image analysis and image manipulation because each input image pixel can be mapped to one neuron in the first layer of the network. Since neurons are only connected to nearby neurons in the following layer, the structural relationships within the image are preserved through the model layers. …