Promptable concept segmentation (PCS) is a segmentation technology driven by "concept prompts". It supports multiple prompt methods such as semantic description and image exemplars, which can flexibly adapt to different segmentation needs and improve the flexibility and accuracy of model segmentation.
