The segment-anything model (SAM), was introduced as a fundamental model for segmenting images. It was trained using over 1 billion masks from 11 million natural images. The model can perform zero-shot segmentation of images by using various prompts such as masks, boxes, and points. In