Depression Diagnosis Via Social Media

[Algorithm Can Recognize Depression By Looking At Instagram Photos]

One of the things that makes depression a very serious disease is the fact that it is silent in nature. No matter how much it destroys the mental landscape, on the outside, the symptoms aren’t always that obvious.

This is why a majority of depression cases go undiagnosed until the toll is too much and the patient’s entire life starts suffering as a result.

Not to mention, that even if patients are proactive enough to seek psychological counseling, there’s no telling how long it would take before the person in charge of their care is able to detect the level of their depression.

This are the circumstances under which people begin to feel the need for a technological intervention, as in recent times, a lot of diseases have found solutions through machine learning and artificial intelligence programs.

Thankfully for them, there are certain programs, like the Watson AI, that do take advantage of current technological innovations to help diagnose all sorts of diseases.

Taking a cue from them, researches from the University of Vermont and Harvard University have developed a special type of Machine Learning Algorithm that can diagnose depression just by looking at your Instagram Pictures.

The reason they chose social media pictures as a means of diagnosis is because people spend a lot of time and energy into selecting what they feel is the perfect post, filter and look for their images.

This type of effort carries with it a deep insight into the psyche of the person and lets the researchers recognize exactly how the subject is feeling at that particular time.

Since taking a few images is not enough to prove the effects of depression, thereby the researchers had to get creative and develop the above-mentioned algorithm, so that they could analyze hundreds of pictures from dozens of patients.

The algorithm works by evaluating each picture based on their color scheme, how happy or likable the pictures are, as well as other parameters such a hue, saturation, color, contrast among others.

When all this information is tallied together, along with the likes and comments the pictures receive, it becomes easy to diagnose the overall mental condition of the patient, in order to find out if they really do suffer from depression.

To prove the efficacy of this system, researchers tested about 500 workers from Amazon’s Mechanical Turk who were also Instagram users.

They were asked to fill up questionnaires and surveys, along with special permission slips that gave researchers access to their Instagram pictures.

Once all their information was put through the same process, it was found that 70 of them were clinically depressed, which is a pretty staggering number to go undiscovered for such a long time.

The researchers hope to carry out more experiments of this sort, in order to define a standard for this sort of testing, so that eventually it could be used on a large scale for easy depression diagnosis.

Article Details

Source: Arxiv
Category : Medical |
Posted on : 27 Aug 2016 @ 16:23

More in Internet

More In Medical


Copyright © TechnologyWOW All Right Reserved.
Privacy Policy | Terms | Disclaimer