Using the same robust facial recognition technology behind our Demographics model, our Celebrity model was trained to recognize the distinctive features of over 10,000 celebrities and use that knowledge to analyze any uploaded image. From there, it returns probability scores, showing the likelihood that the image contains the face of some of the most well-known A-listers.
Why it works
From red carpet pictures to paparazzi snaps, celebrities are some of the most photographed people on earth and the most well documented. These two things are key to our Celebrity model's success.
While visual recognition doesn't need a lot of examples to learn particular tags, the more high quality, well-labeled data it sees, the better the technology gets at recognizing things. With so many pictures of notable stars, we were able to get multiple examples of the faces we wanted to learn and easily match the relevant tags, i.e., their names, with each.
So, whether you want to tag red carpet pictures or make celebrity comparisons like Popsugar’s Twinning app, this is the perfect model for you. Recent social media buzz, including a shoutout from Ellen Degeneres herself, saw Popsugar’s app go viral, much like our model did when we first released it. Imagine what your app could do.
Ready to build your own? Sign up for your free API key and get started today.
Editor’s note: This article was originally posted on March 29th, 2017 and has been modified to bring it up-to-date.