11 repos
Deep learning systems for automated segmentation of medical images using large-scale foundation models like SAM (Segment Anything Model) and CLIP adapted for clinical workflows. The cluster centers on PyTorch-based implementations for segmenting medical structures across modalities—particularly ultrasound and CT—combining computer vision techniques with domain-specific medical imaging challenges. Repositories here explore both foundational model architecture (MedSAM, MedSAM2, MedCLIP-SAMv2) and practical applications of these models to specialized clinical segmentation tasks (SegVol, adapt_med_seg).