Deep learning microscopy

Abstract

We demonstrate that a deep neural network can significantly improve optical microscopy, enhancing its spatial resolution over a large field-of-view and depth-of-field. After its training, the only input to this network is an image acquired using a regular optical microscope, without any changes to its design. We blindly tested this deep learning approach using various tissue samples that are imaged with low-resolution and wide-field systems, where the network rapidly outputs an image with remarkably better resolution, matching the performance of higher numerical aperture lenses, also significantly surpassing their limited field-of-view and depth-offield. These results are transformative for various fields that use microscopy tools, including e.g., life sciences, where optical microscopy is considered as one of the most widely used and deployed techniques. Beyond such applications, our presented approach is broadly applicable to other imaging modalities, also spanning different parts of the electromagnetic spectrum, and can be used to design computational imagers that get better and better as they continue to image specimen and establish new transformations among different modes of imaging.

Machine Learning for Analysis of Microscopy Images: A Practical Guide

“…ML has been applied with great success in various fields of microscopy, such as brightfield (Rivenson et al, 2017), holographic (Rivenson & Ozcan, 2018), and fluorescence (Wang et al, 2019). It has proven to be particularly useful for tasks like image reconstruction and transformation.…”

Automatic identification of crossovers in cryo‐EM images of murine amyloid protein A fibrils with machine learning

“…However, the methods of statistical learning, in particular CNNs, provide a promising technique with regard to automated image analysis. 2016; Petrich et al, 2017;Rivenson et al, 2017;Furat et al, 2019). 6.…”

“…Due to its fully convolutional nature, the input and output sizes of the network are arbitrary. 2016; Petrich et al, 2017;Rivenson et al, 2017;Furat et al, 2019). Integrating these techniques into the analysis process of fibril images is a promising approach that could drastically reduce the efforts needed to process these images using completely interactive approaches alone.…”

On-chip lens-less holographic trichoscopy / trichometry technologies: microinterferometric, Fourier spectral and correlographic techniques for clinical trichology.

“…Микроскоп / трихоскоп, трёхмерно-реконструирующий и автоматически идентифицирующий в своём программном обес-печении структуры волос, должен быть настоль-ко сложен, чтобы иметь возможность проводить соответствующие типы анализа, вплоть до мик-роколориметрии и псевдоспектрального анализа (в биохимических целях) волос, и настолько прост, чтобы не отпугивать потенциального по- [12] и томографии [13] на чипе с автоматической идентификацией объектов по данным высокоразрешающей мик-роскопии, получаемых на этих установках [14]. Конструкция этих устройств настолько проста, что их использование и даже изготовление мо-жет быть осуществлено человеком без физиче-ского или инженерного образования.…”