Meta, the Mark Zuckerberg-led parent company behind social media giants such as Facebook and Instagram, has announced the launch of a new artificial intelligence (AI) model capable of identifying particular objects within an image. Meta has also released a dataset of image annotations that it claims is the most extensive of its kind.
The Segment Anything Model (SAM) is an advanced object recognition model that identifies objects within images and videos even if it has not encountered them during its training phase. Users can select objects by clicking on them or by using text prompts, such as the word "cat" or "chair" and so on. In a demonstration, SAM was able to draw boxes around multiple cats in a photo accurately in response to the written prompt.
ALSO READ: Meta Looking To Reduce Remote Working: CEO Mark Zuckerberg
Meta has been utilizing technology comparable to SAM internally for duties such as tagging photos, moderating prohibited content, and recommending posts to Facebook and Instagram users. The release of SAM will expand access to this type of advanced technology beyond the company's internal operations.
Meta has made the SAM model and dataset available for download under a non-commercial license. However, users who upload their images to the prototype must agree to use the tool only for research purposes.
ALSO READ: Meta Verified Waitlist Opens In India. All You Need To Know
Meta said in a blog post that SAM has the potential to be used in several domains that require finding and segmenting any object in any image. In the future, SAM could become a component in larger AI systems for a more general multimodal understanding of the world, such as understanding a webpage's visual and text content. SAM could also be utilized in the AR/VR domain, enabling users to select an object based on their gaze and then "lifting" it into 3D.
The company has suggested that SAM could have various applications for content creators, including the ability to extract image regions for use in collages or video editing. Additionally, the model could prove useful in scientific research, allowing scientists to locate and track animals or objects of interest within video footage of natural occurrences on Earth or in space.