New Meta AI algorithm for object recognition in photos - images

Meta's ambitions in the AI space have been known for years, but efforts are intensifying after the latest developments of its competitors (see Microsoft, OpenAI, Google). The company went on to launch the first demo version of its SAM (Segment Anything Model) tool, which has the ability to detect, identify and segment objects through photos and images, which it then processes in various ways.

The difference between SAM and similar AI tools has to do with the fact that it is based on a huge database that allows it to detect with unprecedented accuracy even the smallest object in an image, and even based on the written commands given by the user. For example, consider an image that includes LEGO characters from various sets. SAM can specifically identify only those that came out of the Lord of the Rings set with one click!

After the image processing - analysis is complete, Meta's system cuts the requested object from the rest and offers the user options to take further actions, such as providing information about the specific object.

Most importantly, Meta has made SAM an open-source tool, so anyone has access to it and can enhance the database with more material, while paving the way for developers to integrate it into various applications, even augmented or virtual reality.