Convolutional neural networks and deep learning can boost the capabilities of standard optical microscopes to levels comparable to those of some higher-end imaging systems.
[Illustration by A. Ozcan]
Deep learning, particularly using convolutional neural networks (CNNs), is transforming a range of disciplines and eclipsing the state of the art achieved by earlier machine-learning techniques. In machine vision, for example, the deep-learning revolution has driven new capabilities in autonomous vehicles, fault analysis, security applications, entertainment and the Industrial Internet of Things. Deep-learning-enabled breakthroughs in voice recognition and speech translation are transforming how we communicate with each other and with our devices. And supervised deep-learning approaches—in which a system learns to classify or otherwise interpret information by analyzing “training sets” of labeled data—have found particular use in biomedicine and medical imaging: disease diagnosis through classification of histological images; determining tumor margins in cancer cases; cell classification and counting; screening patients for certain eye diseases using optical coherence tomography scans; and a host of other areas.
…