Sam2 For Medical Images

--- language: en license: apache-2.0 library_name: pytorch tags: - sam2 - segment-anything - medical-imaging - image-segmentation - finetuned --- # SAM2 Medical Fine-tuned Model This is a fine-tuned version of Meta's Segment Anything Model 2 (SAM2) optimized for medical image segmentation tasks. ## Model Details - **Model ID**: drippyone/sam2-for-medical-images - **Base Model**: sam2_hiera_base - **Fine-tuned Dataset**: Kvasir-SEG (medical endoscopy images) - **Task**: Image Segmentation - **Framework**: PyTorch - **Upload Date**: 2026-05-26 ## Intended Use This model is designed for medical image segmentation, particularly for segmenting lesions and anatomical structures in endoscopic images. ## Training Details - Dataset: Kvasir-SEG medical imaging dataset - Task: Fine-tuned for medical image segmentation - Architecture: SAM2 Hiera Base ## How to Use ```python import torch from sam2.build_sam import build_sam2 # Load the fine-tuned model device = torch.device("cuda" if torch.cuda.is_available() else "cpu") model = build_sam2(checkpoint="sam2_medical_finetuned.pth") model = model.to(device) # Use for inference with torch.no_grad(): # Your inference code here pass ``` ## Limitations - Model is specifically optimized for medical imaging tasks - Performance may vary on non-medical datasets - Requires appropriate medical imaging data for optimal results ## Citation If you use this model, please cite the original SAM2 paper and acknowledge the medical imaging dataset used for fine-tuning. ## License This model is licensed under the Apache 2.0 License - same as the original SAM2 model.

Source context: 0 downloads · 0 likes · Pipeline image-segmentation · Library pytorch · Repo drippyone/sam2-for-medical-images