Meta’s Incoming AI Photo-Editing Tool Auto-Selects & Erases Objects With A Click
Get ready to say goodbye to the unsightly photobombers and pesky clouds ruining your picture-perfect moments, thanks to Meta’s latest artificial intelligent offering, the ‘Segment Anything Model’ (SAM).
Trained on an enormous dataset of over one billion images, SAM can analyze every object in a photo down to the tiniest detail. During a demo, it identified every fruit and vegetable thrown together in a crate.
Separating them was as simple as clicking or drawing a box over a carrot or capsicum. With the highlighted objects in view, erasing them only took a single click. Not only that but SAM is also being developed to work seamlessly with videos.
SAM’s unique and efficient segmentation model distinguishes it from similar models, requiring individual guidance for each step. Moreover, many different models are trained on large amounts of manually annotated objects, which takes up much time and resources.
Video via Meta
Meta predicts that SAM’s capabilities will be in high demand among professions that rely on rapid image-based results, such as science photographers and journalists. It also envisions the self-driving and robotics industries benefiting from SAM’s technology.
The program is currently only available for demo tryouts as of now.
Today we're releasing the Segment Anything Model (SAM) — a step toward the first foundation model for image segmentation.
— Meta AI (@MetaAI) April 5, 2023
SAM is capable of one-click segmentation of any object from any photo or video + zero-shot transfer to other segmentation tasks âž¡ï¸ https://t.co/qYUoePrWVi pic.twitter.com/zX4Rxb5Yfo
