Smartphone Takes Microscopic Images

Image of a blood smear from a cell phone camera (left), following enhancement by the algorithm (center), and (right) taken by a lab microscope. (Source: Ozcan Group, UCLA)

Researchers at the UCLA Samueli School of Engi­neering have demonstrated that deep learning, a powerful form of arti­ficial intelligence, can discern and enhance micro­scopic details in photos taken by smart­phones. The technique improves the reso­lution and color details of smart­phone images so much that they approach the quality of images from labora­tory-grade micro­scopes. The advance could help bring high-quality medical diagnostics into resource-poor regions, where people other­wise do not have access to high-end diagnostic techno­logies. And the technique uses attach­ments that can be inex­pensively produced with a 3-D printer, at less than $100 a piece, versus the thousands of dollars it would cost to buy labora­tory-grade equipment that produces images of similar quality.

Cameras on today’s smart­phones are designed to photo­graph people and scenery, not to produce high-reso­lution micro­scopic images. So the researchers developed an attach­ment that can be placed over the smart­phone lens to increase the reso­lution and the visi­bility of tiny details of the images they take, down to a scale of approxi­mately one millionth of a meter. But that only solved part of the challenge, because no attach­ment would be enough to compen­sate for the difference in quality between smart­phone cameras’ image sensors and lenses and those of high-end lab equipment. The new technique compen­sates for the difference by using arti­ficial intelli­gence to reproduce the level of reso­lution and color details needed for a labora­tory analysis.

The research was led by Aydogan Ozcan and Yair Rivenson. Ozcan’s research group has intr­oduced several inno­vations in mobile micro­scopy and sensing, and it maintains a particular focus on deve­loping field-portable medical diag­nostics and sensors for resource-poor areas. “Using deep learning, we set out to bridge the gap in image quality between inex­pensive mobile phone-based micro­scopes and gold-standard bench-top micro­scopes that use high-end lenses,” Ozcan said. “We believe that our approach is broadly applicable to other low-cost micro­scopy systems that use, for example, inex­pensive lenses or cameras, and could facilitate the replacement of high-end bench-top micro­scopes with cost-effective, mobile alter­natives.” He added that the new technique could find numerous appli­cations in global health, tele­medicine and diag­nostics-related appli­cations.

The researchers shot images of lung tissue samples, blood and Pap smears, first using a standard labora­tory-grade micro­scope, and then with a smart­phone with the 3D-printed micro­scope attach­ment. The researchers then fed the pairs of correspon­ding images into a computer system that “learns” how to rapidly enhance the mobile phone images. The process relies on a deep-learning–based computer code, which was developed by the UCLA researchers. To see if their technique would work on other types of lower-quality images, the researchers used deep learning to success­fully perform similar trans­formations with images that had lost some detail because they were com­pressed for either faster trans­mission over a computer network or more effi­cient storage. (Source: UCLA)

Reference: Y. Rivenson et al.: Deep Learning Enhanced Mobile-Phone Microscopy, ACS Phot., online 15 March 2018; DOI: 10.1021/acsphotonics.8b00146

Link: California NanoSystems Institute, University of California, Los Angeles, USA

Speak Your Mind

*