01 October 2016
It’s odd how a topic for a blog can creep up on you. This week I listened to a BBC Radio 4 “Four Thought” podcast talking about depression and the healing power of poetry (the episode is called Healing Minds and was first broadcast on 30th March 2016) and in the same week came across an interesting photography related article in the MIT Technology Review called “How an Algorithm Learned to Identify Depressed Individuals by Studying Their Instagram Photos”.
Researchers have found that people suffering from depression prefer darker colours. I’m not sure how long ago that link was made, but in the present day with “big data” and powerful computers it raises the possibility of diagnosing depression en masse by analysing the photos people post to social-media sites such as Instagram.
The work of US based computer scientists Andrew Reece at Harvard University and Chris Danforth at the University of Vermont has found significant correlations between the colours in photos posted to Instagram and an individual’s mental health.
The researchers surveyed 170 people. Around 70 of them were clinically depressed; the survey asked various additional questions about their condition, such as the original date of their diagnosis. The researchers then invited the sample group to share their Instagram posts for the study. A sample of 170 is not large but their Instagram downloads resulted in a database of over 40,000 photographs which were analysed. For each healthy user, the researchers chose the 100 most recent photographs to be rated. For depressed individuals, the researchers chose the 100 photographs posted before their diagnosis.
(With some help) each image was judged to say how interesting, likeable, happy and sad it was. The researchers also evaluated the images using objective measures such as the average hue, colour saturation, contrast, etc. together with counting the number of faces in each image, on the assumption that faces are a proxy for an individual’s level of social activity.
Armed with this data, the researchers used a computer to spot correlations between depression and image properties. They found that depressed individuals post images that are bluer, greyer, and darker, and receive fewer “likes”, than those posted by heathy individuals.
To determine how well the algorithm can identify depressed individuals using images they post on Instagram the researchers analysed the images posted by 100 individuals and found that the algorithm correctly identified 70 percent of those who were depressed. The article said that’s a higher success rate than achieved by GPs evaluating patients (I’m not sure how they measured that). Nevertheless that mental illness can be accurately detected earlier, allowing for more effective intervention, is a great hope for future use of technology in this area.
The researchers found strong links regarding Instagram filters used: Depressed individuals had a clear favourite, the “Inkwell” filter which converts colour images to black-and-white. By contrast, healthy individuals preferred the filter called “Valencia”, which lightens images.
I wonder what that makes of those of us who make black and white images via more traditional means? Moreover I wonder what my Instagram account says about me…
If you have any images to the issue of depression please upload to a gallery and share the details in the comments below.