Have you tried the Face App?
if you do social media you will have seen the collages produced by the age changing, gender bending app. It recognises the key points of your face then applies the formula values associated with age, youth , male or female genders and even a "good loooking" formula.
The science behind it is the one which means your passport photo should be without smiling and glare free on the glasses. There's been a lot of money spent on facial recognition.
There's also been a lot of money on machine learning so what happens if you put the two together?
In the world of medicine the answer is Face2gene .
This app has learned the typical features of people with multiple genetic defects. Unlike a human geneticist who must study books, recall the details and make a judgement call before ordering genetics, the app compares and produces a likelihood match for hundreds of conditions.
Clearly there are caveats in using this app.
Its designed for clinicians. It's designed as a diagnostic aid not a replacement for clinical expertise.
Here's an example.
I have a high gestalt match for Gapo syndrome. Never heard of it? Me neither. The gestalt match means my photo matches the aggregated "deep learning" of the system.
I do not have Gapo syndrome, I have hair, relatively normal intellligence, and have made it to over 50.
The system is essentially providing the clinician with a memory which never forgets, never struggles to match the pattern. As the ultimate genetic aide memoire it's got a bright future, as a replacement for a clinical geneticist it's a dodo.