top of page

Artificial intelligence detects rare diseases

Global Goals & Global Society
Artificial intelligence detects rare diseases

The free Face2gene app uses artificial intelligence and facial recognition to diagnose rare diseases in children.

Facial recognition to diagnose diseases

Several rare diseases leave particular marks and traces on the face of children, doctors often diagnose these diseases by analyzing the facial features, so the idea of this application was born.

In 2014 the start-up FDNA launched the Face2gene application, it can detect rare diseases by analyzing the facial features of the child. This application can recognize today more than 1000 genetic diseases.

The algorithm that detects diseases, GestaltMatcher, uses 17560 photographs of people with rare genetic diseases to model certain characteristics present on the face of children with the disease.

Doctors and geneticists are increasingly using this medical software to facilitate their research and diagnosis. "This is an important contribution for very rare diseases, where only a few patients have been reported in the world," say the start-up's researchers. The app supports early detection of disease and rapid prescription of treatment.

Areas for improvement

However, the software cannot detect the majority of genetic diseases, because they are rare! You need 7 photos of patients with the disease for it to be recognized by the algorithm, and getting 7 photos of patients with rare disease is not easy.

The application is problematic on another level: it is more effective on Caucasian faces than on faces of other ethnic groups. Researchers and programmers are trying to reduce this difference in detection by doing more research and adding more photos to their database. This problem is found in other types of technological devices, some brands of soap dispensers do not detect black hands, or artificial intelligence systems have a harder time detecting skin cancers in black people. This illustrates the need to make artificial intelligence more inclusive and less discriminatory in new technologies.

Future updates will allow the application to increase their photo quantity, detection reliability, the number of rare genetic diseases detectable by the software and thus diagnose more and more diseases in a very simple way.

Apps like these, with the help of increasing the network of global society, are well on their way to being an asset to the well-being of our society in the future. Like all innovations, this one still needs some test runs, the patience and belief in its idea and the support of various funding measures to support scientists with great ideas and to achieve certain Sustainable Development Goals.

More information:


bottom of page