South Korean startup Lunit Inc. has come under the spotlight after it unveiled a deep learning based digital image analysis solution that is expected to quickly replace the conventional medical diagnosis system with more accurate and efficient disease detection capability.
Lunit, established by a team of young Korean engineers in 2013, has developed a deep learning based image recognition system for medical image analysis. Deep learning is an algorithm and data-based machine learning that allows a computer to learn and evolve on its own. Using the deep learning mechanism, Lunit’s image recognition system scans and analyzes x-ray pictures and computerized tomography (CT) pictures to provide medical interpretation of the scanned images.
As an existing medical diagnosis system based on x-ray and CT scanned images of body relies heavily on a medical doctor’s judgment, patients sometimes get different analyses of the same scanned image from different doctors. But the Lunit’s digital image recognition system studies scanned images in just five seconds and provides analysis with an accuracy rate of 95 percent based on the intelligence it had acquired based on the study of over 100,000 x-ray pictures, said Anthony S. Paek, co-founder and CEO of Lunit.
With this cutting-edge image recognition system, Lunit ranked the seventh at the internationally-renowned ImageNet Large Scale Visual Recognition Challenge for the first time among Korean participants in 2014, and stepped up to the fifth in 2015 ahead of Google that ranked sixth at the competition on digital image recognition technology.
Lunit has differentiated its technology by concentrating on developing a medical image recognition system, especially for tuberculosis and cancer detection. Located in Korea with a high number of university-run big, general hospitals, it is easier for the company to collect relatively more medical cases compared to its rivals in other countries.
The venture claims its system also allows physicians to detect a patient’s disease faster than its competitors’ devices. Even more, its medical x-ray picture analysis system enables to discover illness from remote places without medical specialists, the company said.
Recognized with its advanced technology, Lunit has attracted investment amounted 2.6 billion won ($2.3 million) through international startup incubation programs including Korea’s Small and Medium Business Administration’s TIPS program, SoftBank Ventures, and Formation 8 in just three years since its establishment.
By Chung Soon-woo
[ⓒ Pulse by Maeil Business News Korea & mk.co.kr, All rights reserved]