medsam

Desktop image annotation tool for segmentation masks, with Segment Anything (SAM), text-to-mask and active learning. Runs offline: your images never leave your computer.

Parameter-efficient brain lesion segmentation using MedSAM and LoRA on the MICCAI WMH Challenge dataset.

VGG16-UNet vs LoRA-fine-tuned MedSAM for prostate segmentation on micro-ultrasound

Anatomy-guided Cross-modal Fusion for automated radiology report generation from DICOM CT scans. Built with CT-CLIP, MedSAM, GatorTron, and LLaMA-3 (LoRA).

MedSAM enhanced with CLIP text descriptors for multi-organ medical image segmentation, evaluated on 24k labeled CT slices from FLARE 2022.

Parameter-efficient LoRA fine-tuning of MedSAM (ViT-B) for prostate segmentation, achieving 95.7% Dice and 91.9% IoU.

Fine-tuning MedSAM and SAM for medical image segmentation with custom prompt strategies

VGG16-UNet vs LoRA-fine-tuned MedSAM for prostate segmentation on micro-ultrasound

Web-based semi-automated retinal fluid segmentation tool for OCT images. Combines traditional image processing with deep learning models (U-Net, MedSAM) to help medical professionals efficiently annotate and segment retinal fluid regions. Features interactive annotation tools, multiple segmentation algorithms, and user-friendly web interface.

Reproducible, model-agnostic prompt evaluation for medical segmentation.

Reproducible cardiac MRI SAX segmentation workflow with ACDC validation, Dice/HD95 metrics, QC overlays, and failure analysis

SegMed: Implementation of MedSAM segmentation and enhancement with FeatUp and GAFL

Machine Learning for Medical Image Processing. Project done for Charité, the university hospital of Berlin. The aim was to segment coronary arteries and then extract a graph from it, in order to aid detection of coronary artery disease (CAD).

This is the official repository of Team Anastasia, which achieved 1st place in ImageCLEF 2025 Dermatological Segmentation Challenge (Subtask 1)..

DICOM ROI annotation & analysis desktop tool (PySide6): SAM / MedSAM / SAM-Med2D-3D interactive segmentation + 5 classical CV ROI detectors, with an MCP bridge for AI-agent driven workflows. 医学影像 ROI 标注与分析桌面工具。

Semi-supervised tumor segmentation with MedSAM2 confidence-gated pseudo-label refinement(KiTS21 & LiTS)

Undergraduate coursework project exploring the MedSAM paper as part of the Machine Learning and Related Applications module.

Benchmarking SAM and MedSAM on ISIC 2018 melanoma segmentation — with a focus on the deployment gap: published SAM papers measure performance using ground-truth-derived prompts that are unavailable in real clinical settings. This project quantifies that gap.

Fine-tune MedSAM with LoRA for parameter-efficient prostate segmentation on micro-ultrasound, underpinned by a VGG16-UNet CNN baseline.

Monorepo containing the frontend (Next.js), backend (Node.js + Express.js), and GPU inference service (TensorRT-optimized Python engine) for a cardiac segmentation application, streamlining local deployment.