

Given the sharp surge in the volume of deep learning articles published in medical journals in 2017 that is commensurate with the trend of growing awareness and interest in deep learning within the radiologic community, the time appears optimal for presenting a guide on deep learning for radiologists that includes a general framework of deep learning research and its applications in the field of radiology.

Other reviews, such as the article by Litjens et al on studies published until February of 2017 ( 6), have presented a comprehensive survey of the literature with an overview of deep learning techniques and applications ( 6, 11– 13). Some have focused primarily on deep learning methodology ( 5, 9, 10). Several deep learning reviews have been published in the last few years. This is particularly important to the field of radiology, with its visual-based data ( 6– 8). Today, CNN is considered to represent the state of the art in image analysis ( 5, 6).Ĭomprehensive academic research, as well as start-up endeavors, is working on finding deep learning solutions that can be applicable to the medical world. In this competition, Krizhevsky and Hinton ( 4) successfully developed a CNN named AlexNet that surpassed other competing classic machine learning techniques. Deep learning–based methods, however, did not receive wide acknowledgment until 2012, in the ImageNet challenge for the classification of more than a million images into 1000 classes. A major breakthrough in the field of deep learning was presented by Lecun and colleagues in 1998 ( 3), whereby they applied their novel convolutional neural network (CNN), LeNet, to handwritten digit classification. Several image analysis models were developed, and the latest advancement in this field is a technique called deep learning.ĭeep learning is considered by some to be an integral part of the Fourth Industrial Revolution ( 2). In the past few decades, developments in medical imaging technology and the increased role of imaging within the diagnostic process have resulted in a rapid expansion of recorded medical visual data, generating a need for novel computational models ( 1).
