A rare hereditary disease is difficult to diagnose. Sufferers have to go through innumerable tests,that take up valuable time of doctors and patients alike with still no definitive result. This additional time delays therapy that could actually be directed to forestall damage. Professors from the University Hospital Bonn and a team of researchers had data of 679 patients suffering from 105 different diseases .They trained a neural network model with 30,000 portrait pictures of affected people and demonstrated how artificial intelligence can be used to perform efficient and reliable facial analysis in cases where the facial structure does show mutations when afflicted with a rare disease. Together with other symptoms and genetic data,it was possible to get accurate diagnosis.

Read more at: https://www.sciencedaily.com/releases/2019/06/190606133805.htm