Deep Learning in medicine is one of the most rapidly and new developing fields of science. Currently, almost every device intended for medical imaging has a more or less extended image and signal analysis and processing module which can use deep learning. It provides quantitative data necessary to make a diagnosis. The obtained quantitative features must be independent of the inter-subject variability and the type of medical device and, above all, must allow for reproducible results in the presence of high noise. The proposed deep learning algorithms should also ensure the independence of the results obtained by the operator of the imaging device and, to be more exact, its position relative to the patient or the parameter settings in the device. In addition, the proposed deep learning algorithms must be tailored for the diagnosis of a specific disease entity. On the other hand, they must allow for reproducible results for high inter-subject variability. These criteria make it difficult to propose a methodology for the deep learning algorithms. This special issue of BioMedical Engineering Online is dedicated to this area of knowledge.
Topics:
- Deep neural network in medical image processing (RTG, USG, CT, PET, OCT and others)
- New deep neural network architecture
- The use of applications with deep machine learning for recognizing objects in a 3D scene
- Deep machine learning in large data sets
- Deep robot learning
- Data mining with deep learning in bioinformatics
- Applications, algorithms and tools directly related to deep learning
Edited by Robert Koprowski