Project link: https://github.com/IDEA-Research/Grounded-Segment-Anything
Meta recently released an AI model called Segment Everything, or SAM for short. SAM can generate masks for any object in any image or video, even objects and image types not encountered during training. SAMs are general enough to cover a wide range of use cases and can be used out-of-the-box in new image domains without additional training. This idea has attracted the attention of many people.
The release of this model caused a sensation in the field of computer vision, indicating that CV will also move towards a unified path. Maybe everyone had a hunch about this, but they didn't expect this day to come so soon. What is faster than the iteration of the basic model is the speed of "second creation" in the research community. Just two days after the paper was published, several domestic engineers came up with new ideas based on this and put them into practice, forming a visual workflow model that can not only divide everything , but also detect everything and generate everything .
