This Machine Uses Your Instagram Photos To Diagnose Depression, Other
Researchers have created a machine that checks Instagram photos to diagnose depression and other mental illnesses.
Social media can give insights into a person's mental state and personal life. A new study conducted by Andrew Reece of Harvard University and Chris Danforth of the University of Vermont suggests that there is a connection between usage of color and the mental state of a person.
The researchers believe that people who post grayer or darker colored images on Instagram are more likely to be depressed in comparison to those who post images of vibrant colors. The researchers have also developed a machine that spots depression in people by identifying the images posted by individuals.
One image can have different interpretations. An image with bright colors can be doctored with an Instagram feature called Inkwell that converts colored photos into a black and white image. Study reports suggest that people who are depressed use Inkwell more often when compared to people who are not depressed.
For the purpose of the study, researchers examined about 170 workers from Amazon's Mechanical Turk service who had Instagram accounts. All the subjects were asked to complete a questionnaire, which also included a standard clinical depression survey. The study subjects were also asked to share pictures from their Instagram account.
The researchers selected about 100 photographs from each subject and asked people to rate them on a scale of 0 to 5 based on how interesting, sad or happy the photos looked. The study also categorized the photos based on saturation, hue and the number of faces each photo had.
The researchers used the photos in a machine-learning algorithm that spotted the correlations between image properties and depression.
The algorithm found that decreased saturation and brightness, and increased hue predicted depression. The study also found the depressed and non-depressed people used filters differently. For instance, the researchers found that depressed people were less likely to use filters.
The researchers found that the algorithm was able to identify depression with a 70 percent success rate. It is not possible for any technique to identify depression with a very high success rate. Medical professionals as well as clinical questionnaires are also not 100 percent accurate.
However, the study found similarities between people who were depressed and the way they posted photos on Instagram.
"More generally, these findings support the notion that major changes in individual psychology are transmitted in social media use, and can be identified via computational methods," report the researchers.
Danforth and Reece suggest that their technique will be helpful in understanding mental illness in people and detecting depression at an early stage, allowing for effective diagnosis.
No comments