20 Exceptional Photos Using Computer Vision to Pick Exceptional Portraits
EyeEm is teaching the computer to sift through millions of photos and find the best subjects.
People are one of the most popular subjects on any photo platform: Portraits are as much fun to take as they are to look at, and a great portrait doesn’t just show a person but reveals something about their personality.
But how do you find the best ones? With millions of photos being uploaded to the internet every day, it’s become impossible for humans to see even a small percentile of them – let alone sort through the photos and find the best ones.
In the future, we’ll have to rely increasingly much on technology to make sense of all the data on the web and to do the heavy lifting for us. Photo platform EyeEm is building a computer vision engine that already scans all of the images uploaded by their users and automatically ranks them. The technology uses deep learning technology to imitate the taste of its photo editors, who train it with pre-curated image collections. Using the data from those photos, it can then teach itself which characteristics make up a great picture.
At the end of 2016 they were able to tally up the numbers and generate a list of the best portraits taken during the year – as curated by computer vision.