An open-source pipeline combining Grounding DINO and Segment Anything to detect and isolate visual objects via text.