Facebook introduced its AI model SAM, can identify different objects from photos.
Facebook’s parent company Meta has launched its artificial intelligence model SAM. This AI model can identify objects by selecting them individually from an image. Meta is already using this technology on its social media platforms Facebook and Instagram. Meta uses this technology to tag photos, ban prohibited content, recommend users’ posts. This project from Meta is now available for download under a noncommercial license.
Meta’s AI model SAM
The research division of Meta has given information about the AI model SAM (Segment Anything Model) in a blog post. The post read, “We are introducing the Segment Anything (SA) project, which will serve as a new task, model, and dataset for image segmentation.
Using the data collection loop of this model, we have generated the largest segmentation dataset to date. In a blog post, Meta’s research division has told that with the help of AI model, it can click on the object of an image and separate it. Or by writing any text, users can separate the object from the image.
Example: If there is a cat in an image, users can click on the cat to separate the cat from the image. Or else they have to type CAT or cat to separate the cat from the image. This way they can separate the cat from the image.
Meta is experimenting with this model with generative AI, which has the ability to generate new content rather than just recognizing or categorizing data. Mark Zuckerberg said it is a priority for him to incorporate such technology into Meta’s apps this year.
