A new study has revealed about a new artificial intelligence technology which can detect some category of genetic disorders by just using a photo of a face. Th study was published in journal Nature Medicine on Monday.
Dubbed as DeepGestalt, the computer program and has been trained on a series of data set of over 17,000 photos of patients affected by more than 200 rare disorders. After that, the program identifies the patterns of markers by using deep learning. The patterns of markers are linked to hundreds of various genetic syndromes.
According to the researchers, while conducting a test with 502 images, more than 90 percent of the time, the computer software successfully placed the correct syndrome in its top 10 list. The program proved to be better than doctors as it detected Angelman syndrome and Cornelia de Lange syndrome in patients.
The images which were used during the test were from patients already diagnosed with a disorder; the software identified possible disorders that have already been diagnosed and did not identify whether each patient had a genetic disorder.
The new development has the potential to help doctors to search for specific genetic markers by combining itself with genome testing. It will cut down time, cost and will remove the emotional burden of the “diagnostic odyssey” on which millions of families embark each year, seeking care for someone with a rare genetic syndrome.