VitaminPScope — AI Digital Pathology Viewer for Whole-Slide Images (WSI).

VitaminPScope is an AI-powered digital pathology platform for viewing whole-slide images (WSI) and performing whole-cell segmentation using deep learning.

It supports SVS, OME-TIFF, and multi-channel microscopy, and runs locally with Docker in a single command.

🚀 Run in one command — no setup, no configuration

🐳 Docker • 🐛 Issues

---

---

✨ Overview

VitaminPScope is a production-ready, web-based platform for:

• Whole-slide image (WSI) viewing
• Multi-channel microscopy visualization
• Slide comparison and annotation
• AI-powered whole-cell segmentation with VitaminP

Built with React + FastAPI + OpenSeadragon + Viv — optimized for real-world deployment.

VitaminPScope is an open-source alternative to commercial digital pathology platforms (e.g., QuPath, HALO) with built-in AI segmentation

---

💡 Why VitaminPScope

• 🧠 Built-in AI — no external pipelines required
• ⚡ Runs locally with Docker in minutes
• 🔬 Designed for real pathology workflows
• 📊 Scales to gigapixel whole-slide images

---

🎬 Demo

---

🔍 Keywords

digital pathology, WSI viewer, whole-slide image viewer, pathology AI, cell segmentation, histopathology, H&E analysis, multiplex imaging, computational pathology, FastAPI, React viewer

---

🚀 Quick Start

1. Clone repository

git clone https://github.com/idso-fa1-pathology/VitaminPScope.git

2. Move into project

cd VitaminPScope

3. Launch the app

⚙️ GPU Support (Optional)

VitaminPScope runs out-of-the-box on both CPU and GPU environments.

🖥️ Run on CPU (default)

docker compose -f docker-compose.public.yml up -d

---

🚀 Run with GPU (Linux + NVIDIA only)

docker compose -f docker-compose.public.yml --profile gpu up -d

---

⚡ One-line setup

🖥️ CPU (default)

git clone https://github.com/idso-fa1-pathology/VitaminPScope.git && cd VitaminPScope && docker compose -f docker-compose.public.yml up -d

🚀 GPU (Linux + NVIDIA)

git clone https://github.com/idso-fa1-pathology/VitaminPScope.git && cd VitaminPScope && docker compose -f docker-compose.public.yml --profile gpu up -d

---

🌐 Open the app

Once running, open in your browser:

👉 http://localhost:3000

---

📂 Data

Place your slides in:

./data/

Supported formats:

• SVS
• NDPI
• TIFF / OME-TIFF

---

🤖 Powered by VitaminP

This platform integrates VitaminP, a state-of-the-art AI model for whole-cell segmentation.

➡️ https://github.com/idso-fa1-pathology/vitaminp

---

🧪 Use Cases

• Histopathology research (H&E analysis)
• Spatial transcriptomics validation
• Multiplex imaging visualization
• AI-assisted cell segmentation workflows
• Digital pathology education and demos

---

🏗 Architecture

• Frontend: React + OpenSeadragon (WSI rendering)
• Backend: FastAPI (API + orchestration)
• AI Service: VitaminP inference engine
• Deployment: Docker Compose (multi-service)

---

🔌 Extensibility

VitaminPScope can be extended via API or integrated into custom pipelines for automated pathology workflows.

---

🧠 AI Model

The VitaminP model is already included inside the Docker image.

✅ No downloads
✅ No configuration
✅ Ready out of the box

---

📄 Associated Paper

VitaminPScope is powered by VitaminP, a cross-modal learning framework for whole-cell segmentation directly from routine H&E histology.

📄 Read the preprint: https://arxiv.org/abs/2604.23799

If you use VitaminPScope in your research, please consider citing this work.

---

📄 License

MIT License