• PDF / 871,471 Bytes
• 13 Pages / 595.276 x 790.866 pts Page_size
• 98 Downloads / 252 Views

DOWNLOAD

REPORT

---

Online ISSN 1976-3786 Print ISSN 0253-6269

REVIEW

Applications of deep learning for the analysis of medical data Hyun-Jong Jang1 • Kyung-Ok Cho2

Received: 28 December 2018 / Accepted: 20 May 2019 / Published online: 28 May 2019 Ó The Pharmaceutical Society of Korea 2019

Abstract Over the past decade, deep learning has demonstrated superior performances in solving many problems in various fields of medicine compared with other machine learning methods. To understand how deep learning has surpassed traditional machine learning techniques, in this review, we briefly explore the basic learning algorithms underlying deep learning. In addition, the procedures for building deep learning-based classifiers for seizure electroencephalograms and gastric tissue slides are described as examples to demonstrate the simplicity and effectiveness of deep learning applications. Finally, we review the clinical applications of deep learning in radiology, pathology, and drug discovery, where deep learning has been actively adopted. Considering the great advantages of deep learning techniques, deep learning will be increasingly and widely utilized in a wide variety of different areas in medicine in the coming decades. Keywords Artificial intelligence  Deep neural networks  Drug discovery  Medical image analysis

& Kyung-Ok Cho [email protected] 1

Department of Physiology, Department of Biomedicine & Health Sciences, Catholic Neuroscience Institute, College of Medicine, The Catholic University of Korea, Seoul 06591, South Korea

2

Department of Pharmacology, Department of Biomedicine & Health Sciences, Catholic Neuroscience Institute, Institute of Aging and Metabolic Diseases, College of Medicine, The Catholic University of Korea, 222 Banpo-Daero, Seocho-Gu, Seoul 06591, South Korea

123

Introduction Machine learning is a field of study which has amassed a collection of various methods that can automatically detect common or even subtle patterns hidden in datasets using a computer (Rajkomar et al. 2019). Deep learning is a subfield of machine learning, which is commonly characterized by algorithms consisting of multi-layered artificial neural networks, especially those containing more than four layers. Because deep learning can extract universal features in very complex datasets, it has exhibited improved performance in many tasks compared to other machine learning techniques (Lecun et al. 2015). These merits have allowed deep learning to become a new approach that can efficiently solve various problems in medicine (Mamoshina et al. 2016; Miotto et al. 2018; Yu et al. 2018). Recent advances in deep learning have demonstrated that diagnoses of diverse diseases based on the classification of radiologic images and histologic slides has almost surpassed human capabilities (Teare et al. 2017; Veta et al. 2015). Deep learning can also provide an outstanding accuracy on the detection of retinopathy from retinal fundus photographs (Gulshan et al. 2016). With the great promises of deep learning technology, the application of deep learni