Most people uploading their photos to Facebook frequently check how many friends Liked and Commented on their pictures. The social networking giant’s facial recognition system shows you faces in the image that you can tag. That’s probably all you see. But there is a whole different story going on in the background.
Facebook pulls a lot of data from your photos
When you upload a photo, the company’s machine learning algorithms scan every pixel to make sense of objects in the photo and add relevant tags. The tags are based on what the algorithm sees, but none of this is shown to the user. Facebook adds tags based on data like whether it is indoors or outdoors, how many people are there, whether they are sitting or standing, and what other objects are there in the photo.
For instance, if you upload a photo of you and your friends cycling, the system will scan the photo and assign it tags like “people riding bicycles” and “outdoor,” etc. even if you do not provide any image description on your end. These tags are located in the source code of the image link. Most users don’t know where to find them.
Fortunately, there is a Chrome extension called Show Facebook Computer Vision Tags that does all the work for you to reveal the hidden tags. After you have installed the extension, it will show you the tags your image is given when you click on the thumbnail to view the enlarged image on Facebook. Adam Geitgey, the creator of the extension, said he created the extension to inform users about how much data companies can pull from their images.
Facebook uses these tags to make your photos more accessible to blind users who depend on screen readers to understand a photo. If you want to see the tags without installing the Chrome extension, Amit Agarwal of Digital Inspiration has described how to do it.
- Open the enlarged version of the photo on Facebook website
- Right-click on the image and click Inspect to open the Chrome Developer Tools
- Find the Alt attribute of the image tag and you’ll see the tags there
It’s pretty easy. Check out this video by Digital Inspiration for detailed instructions: