By using AI, deep learning, and digital health data, Zebra Medical Vision (the Shefayim-located health care start-up) plans to verify the exotic diagnoses by offering radiologists with high-performance formulas to precisely verify and avoid breast, lung, cardiovascular, liver, and bone diseases.
Prior to establishing the firm, Eyal Gura (Zebra CEO) was the founder of PicScout, a firm purchased in 2011 by Getty Images that designed a platform to verify online images’ digital fingerprint. Gura was motivated to establish Zebra, although, after being hospitalized in Mexico after an accident during scuba diving.
While radiologists all over the world have been employing CAD (computer aided detection) on the basis of conventional image processing equipments for the past decade, the system creates a huge number of fake positive results. The deep learning-supported detection platform by Zebra has already resulted in a deep lowering of fake diagnosis rates in pilot programs and trials abroad. It can slash down undetected cases by over 50% in case of breast cancer.
On a related note, researchers have designed an AI system that can precisely detect small fragments of lung cancer in CT scans that radiologists frequently have a hard time to recognize. The AI system is almost 95% precise, in comparison to 65% when performed by human eyes, scientists claimed.
“We employed the brain as a model to generate our system,” claimed a doctoral candidate in the U.S. at University of Central Florida, Rodney LaLonde, to the media in an interview. The approach is akin to the algorithms that facial-recognition tool employs. It scans millions of faces seeking for a specific pattern to get its match.
The group entered over 1000 CT scans into the program they designed to assist the device learn to seek for the tumors. They had to educate the computer about various things to assist it learn correctly.