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, Chancellor’s Professor of Electrical and Computer Engineering and Bioengineering, and Yair Rivenson, a UCLA postdoctoral scholar. 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.
The Latest on: Deep learning
via Google News
The Latest on: Deep learning
- Delving into Deep Learning on February 6, 2019 at 6:23 am
Deep leaning is one of the most profound but widely recognized phenomenon taking place in the world of Information Technology. That statement opened a talk by Chris Rowen, CEO and co-founder of Babble... […]
- Unity Bringing keynote and Mixed Reality and Deep Learning Talks to GDC 2019 on February 6, 2019 at 5:07 am
Unity upcoming lineup for this year’s GDC 2019 includes a keynote and talks from Unity engineers and researchers on mixed reality, spatial computing, and neural networks. Unity’s keynote is ... […]
- PhaseStain: the digital staining of label-free quantitative phase microscopy images using deep learning on February 5, 2019 at 3:22 pm
Using a deep neural network, we demonstrate a digital staining technique, which we term PhaseStain, to transform the quantitative phase images (QPI) of label-free tissue sections into images that are ... […]
- Deep500: ETH Researchers Introduce New Deep Learning Benchmark for HPC on February 5, 2019 at 12:52 pm
ETH researchers have developed a new deep learning benchmarking environment – Deep500 – they say is “the first distributed and reproducible benchmarking system for deep learning, [and] provides softwa... […]
- Deep Learning — What’s the hype about? on February 5, 2019 at 8:31 am
To say artificial intelligence (AI) is transforming healthcare would be an understatement. Thanks to enormous advancements in computer processing power, as well as the increase of data collection ... […]
- Google Tries to Patent Healthcare Deep Learning, EHR Analytics on February 5, 2019 at 7:05 am
February 05, 2019 - Google has applied for a patent for a deep learning system that aggregates EHR data into a “timeline” in order to predict potential adverse events. The “system and method for predi... […]
- Backed by Benchmark, Blue Hexagon just raised $31 million for its deep learning cybersecurity software on February 5, 2019 at 6:02 am
Nayeem Islam spent nearly 11 years with chipmaker Qualcomm, where he founded its Silicon Valley-based R&D facility, recruited its entire team and oversaw research on all aspects of security ... […]
- Deep Learning 'Godfather' Yoshua Bengio Worries About China's Use of AI on February 4, 2019 at 2:53 pm
An anonymous reader quotes a report from Bloomberg: Yoshua Bengio, a Canadian computer scientist who helped pioneer the techniques underpinning much of the current excitement around artificial intelli... […]
- MIT Deep Learning Basics: Introduction and Overview with TensorFlow on February 4, 2019 at 12:39 pm
MIT Deep Learning series of courses (6.S091, 6.S093, 6.S094). Lecture videos and tutorials are open to all. As part of the MIT Deep Learning series of lectures and GitHub tutorials, we are ... […]
via Bing News