Mahsa Vahedian Aboutorabi - Department of Radiation Sciences, Iran University of Medical Sciences, Tehran, Iran.

Introduction:One of the most promising areas of health innovation is the application of artificial intelligence(AI),primarily in medical imaging. Publications on AI haved rastically increased from about۱۰۰–۱۵۰per year in۲۰۰۷–۲۰۰۸ to۷۰۰–۸۰۰per year in ۲۰۱۶–۲۰۱۷. Magnetic resonance imaging and computed tomography collectively account for more than۵۰%of current articles.Materials and methods: In medical imaging, many publicly available ConvNet models (VGGNet, ResNet, Inception V۳ and DenseNet) have been used thus far۲,۱۱.Kermany et al.showed promising diagnostic applications for deep learning and transfer learning techniques in detection of three major retinal conditions, namely DME, CNV and drusen, from images captured using OCT, a technique that employs a retinal-imaging device that uses infrared light and low-coherence interferometry to scan through the retinal layers. Further validation of the effectiveness of the authors’ deep learning approach for medical diagnoses was conducted on a set of children chest X-rays (CXR) consisting of ۵,۲۳۲ training images from ۵,۸۲۶ patients (۲,۵۳۸ bacterial pneumonia, ۱,۳۴۵ viral pneumonia and ۱,۳۴۹ healthy) and ۶۲۴ images (۲۳۴ healthy and ۳۹۰ pneumonia) from ۶۲۴ patients. They achieved an accuracy of ۹۲.۸%.Result: Radiologists are already familiar with computer-aided detection/diagnosis (CAD) systems, which were first introduced in the ۱۹۶۰s in chest x-ray and mammography applications . However, advances in algorithm development, combined with the ease of access to computational resources, allows AI to be applied in radiological decision-making at a higher functional level. AI will surely impact radiology, and more quickly than other medical fields. It will change radiology practice more than anything since Roentgen. Radiologists can play a leading role in this oncoming.