You can now be diagnosed with depression just by looking at your Instagram
Surely you have more than one melancholic friend on your social networks who misses their partner or who has had a problem and he does nothing but posting deep phrases and photos of sunrises It doesn't take a genius to know that he is depressed However, when feelings and mood are not shown directly, is it possible to know what is happening to that person? Well now there is an algorithm that takes care of studying Instagram profiles to know if a person she's depressed.And the most striking thing is that correct in 70% of cases
This is an algorithm or program developed by researchers from North American universities of Harvard and Vermont Said program is capable of studying the photographs published by a user to extract details of what may be happening. Issues such as the color, the user's face or the periodicity are key to knowing if you are a he althy person or if you suffer from depression.
The algorithm has been tested on 166 profiles of Instagram , thus analyzing a whopping 43,950 photos, according to the study carried out. For this, the color of the images has been studied, identifying the absence of this with tendencies of depressed people.It has also been indicated whether to show the face of the user of the profile, studying with visual recognition the number of people that appear in the photos. In addition, the assiduity in this social network can also be another key to discover a disease, and it is that people with depression tend to publish more content. All this without letting go of the likes and the comments These have been the results:
The algorithm has been successful in 70 percent of the cases A more than remarkable percentage if compared to that of the professionals flesh and blood, which apparently only have a correct response of 42 percent of cases of depression In addition, the study shows that people with depression tend to use the Inkwell filter, which applies a layer of black and white to the image, while the he althy people use the Valencia filter, to give light and color to the scene.
The result also indicates that people with depression tend to show photos with faces, although in fewer numbers than people who are in good he alth mental he alth. In addition, depressed people tend to be quite assiduous in posting photos Indicators that have allowed this algorithm to identify them more easily, skipping only the 17 % of cases and giving 23% false alarms Similarly, posts by people with depression usually get few likes but a large number of comments where interact with other users.
Without a doubt, a tool that can help in the future to identify cases of depression through social networks. A clear indication of what is going through the heads of the people behind these profiles, and who is getting closer to being identified by a robot.
