An artificial neural network is used to enhance the resolution of microscopic images. Source: Ozcan Research Group/UCLAAn artificial neural network is used to enhance the resolution of microscopic images. Source: Ozcan Research Group/UCLA

A machine-learning method known as “deep learning,” which uses multi-layered artificial neural networks to automate data analysis, is a key technology behind advances in applications such as real-time speech recognition and automated image labeling. Now, it is also being used in ways that could aid diagnostic medicine.

The deep learning approach can automatically identify abnormalities in X-rays, CT scans and other medical images and data. New applications reported by researchers at UCLA involve reconstructing a hologram to form a microscopic image of an object, and improving optical microscopy.

“This deep-learning-based framework opens up myriad opportunities to design fundamentally new coherent imaging systems, spanning different parts of the electromagnetic spectrum,” said Aydogan Ozcan, an associate director of the California NanoSystems Institute at UCLA (CNSI).

For one study, published in Light: Science and Applications, the researchers produced holograms of Pap smears, blood samples and breast tissue samples. In each case, the neural network learned to extract and separate the features of the true image of the object from light interference and other physical byproducts of the image reconstruction process.

In an early version of the manuscript accepted for publication, the researchers refer to their approach as “an entirely new framework to conduct holographic imaging by rapidly eliminating twin-image and self-interference-related spatial artifacts.”

For another study, published in the journal Optica, the researchers used the same deep-learning framework to improve the resolution and quality of optical microscopic images. That sort of advance could help diagnosticians or pathologists looking for very small-scale abnormalities in a large tissue or blood sample.

“These results are significant for various fields that use microscopy tools,” the researchers state in their study abstract. “Beyond such applications, the presented approach might be applicable to other imaging modalities.”

In addition to his position at CNSI, Ozcan serves as the Chancellor's Professor of Electrical and Computer Engineering at the UCLA Henry Samueli School of Engineering and Applied Science (HSSEAS). His research is supported in part by the National Science Foundation.