Starting first in the US, the feature would include people, places or things related to the image from the Knowledge Graph's database of billions of facts, helping users explore the topic more.

 

"To generate these links to relevant Knowledge Graph entities, we take what we understand about the image through deep learning, which evaluates an image's visual and text signals, and combine it with Google's understanding of the text on the image's web page," the company said in a statement.

 

This information helps Google determine the most likely people, places or things relevant to a specific image.

 

The company then matches this with existing topics in the Knowledge Graph, and then surface them in Google Images.

 

This feature will start to appear on some images of people, places and things in Google Images and will expand to more images, languages and surfaces over time.