Meta, the parent company of Facebook, has unveiled its Segment Anything Model (SAM) – a powerful artificial intelligence model that can identify individual objects within an image, including those that it has not previously encountered. The company also released what it claims is the largest dataset of image annotations of its kind. “We built the largest segmentation dataset to date (by far), with over 1 billion masks on 11M licensed and privacy respecting images,” the company said in its paper.
In a blog post, Meta’s research division said that SAM can be used to select objects by clicking on them or writing text prompts, such as “cat” to identify all cats in a photograph.
Meta has been experimenting with generative AI, which creates new content rather than simply identifying or categorising data. CEO Mark Zuckerberg has said that incorporating such technology into Meta’s apps is a priority this year. Examples of generative AI tools that the company is developing include one that creates surreal videos from text prompts and another that generates children’s book illustrations from prose.
Meta already uses similar technology internally to tag photos, moderate prohibited content, and recommend posts to users of Facebook and Instagram. The release of SAM is expected to broaden access to this type of technology. The SAM model and dataset are available for download under a non-commercial licence, and users uploading their own images to an accompanying prototype must agree to use it only for research purposes.
The release of SAM comes amid a wave of artificial intelligence breakthroughs from big tech companies. Microsoft-backed OpenAI’s ChatGPT chatbot, which uses generative AI, has been a success and has triggered a rush of investments and a race to dominate the space.