Increasing a microscope’s effective field of view via overlapped imaging and machine learning

Abstract

This work demonstrates a multi-lens microscopic imaging system that overlaps multiple independent fields of view on a single sensor for high-efficiency automated specimen analysis. Automatic detection, classification and counting of various morphological features of interest is now a crucial component of both biomedical research and disease diagnosis. While convolutional neural networks (CNNs) have dramatically improved the accuracy of counting cells and sub-cellular features from acquired digital image data, the overall throughput is still typically hindered by the limited space-bandwidth product (SBP) of conventional microscopes. Here, we show both in simulation and experiment that overlapped imaging and co-designed analysis software can achieve accurate detection of diagnostically-relevant features for several applications, including counting of white blood cells and the malaria parasite, leading to multi-fold increase in detection and processing throughput with minimal reduction in accuracy.

A Systematically Optimized Miniaturized Mesoscope (SOMM) for large-scale calcium imaging in freely moving mice

“…By combining the flexibility of diffractive optics with a powerful optimization strategy, our approach offers a platform suitable for a broader range of other imaging applications such as air quality monitoring [47] and parasite screening [48], where a combination of large FOV, reduced aberrations under mechanical and form constraints are required. As such we anticipate that the proposed method can serve as a guideline for optimizing integrated optical observation devices.…”

The Impact of Artificial Intelligence on Microbial Diagnosis

“…Incorporation of inexpensive mobile devices, such as smartphones and tablets, into the diagnostic process equipped with cameras attached to microscopes facilitate image acquisition and analysis. The integration of mobile technology, aided by the development of specific smartphone applications, has automated and accelerated the diagnostic process [ 110 , 111 , 112 , 113 , 114 , 115 ]. High-resolution cameras and the precision of CNNs has resulted in outstanding classification rates, with some instances achieving results within a mere 10 s [ 113 , 116 ] ( Figure 3 ).…”