Most of us have a Facebook account and a majority of us have uploaded some of our photographs on Facebook at least once. But do you know, whenever you upload any photograph(whether it is a portrait one or a group photograph) facebook very cleverly determines the details about the image?

 

You may not have described the photograph yet Facebook get to know what the picture is all about.

 

Whether you are enjoying with your friends at a birthday party, taking a bike ride, having dinner with family, Facebook can determine accurately what the picture is all about. It can also tell how many numbers of people are inside the photograph without any error. This is really very interesting. Isn’t it?

 

Facebook Computer Vision Tags

 

You must be now very curious about what Facebook actually know about your photograph. Follow these simple steps to know about the data captured by Facebook’s visual recognition algorithms:-

 

1.  Log in to your Facebook account and open any of the photographs of your choice. I have selected a random Donald Trump’s Photograph.

 

2.  Open the photograph and enter the full-screen mode by clicking on the extreme upper right two-headed arrow.

 

 

Facebook photo fullscreen

 

 

3. After entering the full-screen, right-click anywhere and select inspect element to open the developer tool option. You can also do that by pressing Ctrl+Shift+I.

 

 

Inspect element option for the photograph

 

 

4. Check the blue highlighted area to know about the details regarding the picture. Whether you have added any description about the photograph or not, Facebook will tell you the exact details regarding what is inside the photograph.

 

 

Chrome Dev Tool showing image information

 

 

You can see the alt attribute describing the image as seen by Facebook which is “Image may contain: 2 people, people smiling and people standing” which is accurate.

 

Let’s try another picture…

 

 

Salman Khan and 4 other people

 

 

Alt attribute of the photo

 

 

The alt attribute describing the image as seen by Facebook which is “Image may contain: 5 people, people smiling, people standing, shoes and indoor” which is really very accurate.

 

So what Facebook’s computer vision algorithms captured so far?

 

They were:

 

  1. Place where the photograph is taken i.e., indoor or outdoor.

  2. Postures of the people(Standing or sitting).

  3. People’s expression(Smiling, etc.).

  4. Objects present(like spectacles, shoes, etc.).

  5. The number of people.

 

This is really very interesting to know how Facebook know about our photographs even if we have not added any description about the image.

 

Share your thought in the comment section below.