Scientists in the USA developed the algorithms.

In the United States, scientists are saying that they have created machine learning algorithms that can identify depressed people by scanning for “clues” on their Instagram accounts.

According to researchers from the University of Vermont and Harvard University, the algorithm they created has a detection rate of 70 percent, adding that it is “more reliable than the 42 percent success rate of general practice doctors diagnosing depression in a person”.
 
“This points toward a new method for early screening of depression and other emerging mental illnesses,” says Chris Danforth, of the University of Vermont.
 
The research carried out in two stages - first by identifying the clues on the Instagram images thats suggests the user may be depressed and second was by teaching the computer to detect those who are suffering from depression using machine learning algorithms.
 
The scientists gained access to feeds and mental health history of 166 participants. From the they analysed 44,000 images using their from well-established psychology research to look for clues within Instagram images that could reveal the user’s state of mind.
 
Although their research is breakthrough and has great potential, its diagnostics still need some work. Danforth said, “This study is not yet a diagnostic test, not by a long shot. But it is a proof of concept of a new way to help people.”