Apppublic

AbdullahIsaMarkus/veterinary-dicom-mcp-local

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes

App README

🐾 Veterinary DICOM MCP Server

First MCP server for veterinary medical imaging with species-specific DICOM enhancement using CLAHE, adaptive histogram equalization, and AI-powered quality assessment.

🎥 Demo Video

[🔗 Watch the Full Demo on Loom](https://www.loom.com/share/edc57face5614164ac822b110ba76c0e?sid=241e64ef-06fc-4f2e-9c3e-2bf32cb5b162)

See the Veterinary DICOM MCP Server in action with real veterinary X-ray enhancement and analysis.

🎯 Hackathon Track 1: MCP Server Implementation

Transforms veterinary radiology with:

—Species-specific algorithms (canine, feline, equine, bovine)
—Advanced image enhancement (CLAHE, adaptive, contrast stretching)
—AI quality metrics (SSIM, PSNR, entropy analysis)
—Agent-ready prompts for diagnostic assessment

Built with Gradio + scikit-image + veterinary expertise from DIRU (Diagnostic Imaging Research Unit).

🔧 MCP Integration

Add this to your Claude Desktop or MCP client:

json

{
  "mcpServers": {
    "veterinary_dicom": {
      "url": "https://huggingface.co/spaces/Agents-MCP-Hackathon/veterinary-dicom-mcp/gradio_api/mcp/sse"
    }
  }
}

🚀 Features

Species-Specific Enhancement

—Canine: Optimized parameters for dog radiology
—Feline: Cat-specific imaging adjustments
—Equine: Horse anatomy considerations
—Bovine: Cattle imaging optimization

Advanced Algorithms

—CLAHE: Contrast Limited Adaptive Histogram Equalization
—Adaptive: Local contrast enhancement
—Histogram: Global equalization
—Contrast Stretch: Percentile-based improvement
—Gamma Correction: Brightness adjustment

Quality Assessment

—SSIM: Structural Similarity Index
—PSNR: Peak Signal-to-Noise Ratio
—Entropy: Information content analysis
—Edge Density: Structure definition metrics

🏥 Clinical Applications

—Diagnostic Imaging: Enhanced visualization for veterinarians
—Quality Control: Automated image quality assessment
—Research: Standardized enhancement for studies
—Education: Teaching tool for veterinary radiology

🏆 Hackathon Innovation

This project represents the first implementation of:

—Veterinary-specific MCP server
—Species-aware medical image enhancement
—AI-powered diagnostic quality assessment
—Integration of veterinary domain expertise with modern AI tools

🔬 Technical Implementation

Built using:

—Gradio: Web interface and MCP server framework
—scikit-image: Advanced image processing algorithms
—PyDICOM: Medical imaging format support
—NumPy/SciPy: Scientific computing foundation

👥 Team

DIRU - Diagnostic Imaging Research Unit Veterinary Medicine, University of Zurich

Combining veterinary medical expertise with cutting-edge AI technology to advance animal healthcare through improved diagnostic imaging.

📖 Usage

1.Upload a DICOM file or medical image
2.Select animal species (canine, feline, equine, bovine)
3.Choose enhancement method (CLAHE, adaptive, etc.)
4.Specify body region for targeted analysis
5.Receive enhanced image with quality metrics and AI assessment

🎯 MCP Tools Available

—enhance_dicom_image: Species-specific image enhancement
—compare_enhancement_methods: Multi-algorithm comparison
—Automated quality metrics and diagnostic prompts for AI agents

Developed for veterinary medicine with ❤️ and cutting-edge web technology

Gradio Agents & MCP Hackathon 2025 - Track 1 Submission