Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 

README.md

🩺 Medical Image Analysis Tool

An AI-powered radiological diagnostic assistant designed to interpret complex medical imaging documents including X-rays, MRIs, CT scans, and Ultrasounds. This tool provides a structured, professional-grade clinical analysis coupled with patient-friendly explanations.


✨ Features

  • Multi-Modality Support: Analyzes any medical imaging format (JPG, PNG, DICOM-derived images).
  • Structured Clinical Reporting:
    • Modality & Region Identification: Precise anatomical positioning.
    • Key Findings: Systematic highlighting of abnormalities and measurements.
    • Diagnostic Assessment: Confidence-ranked primary and differential diagnoses.
  • Patient-Friendly Logic: Translates clinical jargon into clear, relatable analogies for patients.
  • Research Grounding: Automatically searches the web for the latest medical literature and treatment protocols supporting the findings.
  • Visual Intelligence: Powered by Gemini 2.0 Flash for state-of-the-art vision reasoning.

🛠️ Tech Stack

  • Model: Gemini 2.0 Flash (Vision-capable)
  • Framework: Agno
  • Search Tool: DuckDuckGo Tools
  • Image Processing: PIL (Pillow)
  • UI: Streamlit

🚀 How to Run

  1. Configure API Key: Provide your GOOGLE_API_KEY in the environment.
  2. Launch:
    streamlit run medical.py
  3. Analyze Image:
    • Upload a scan (e.g., a chest X-ray).
    • Clinical analysis will be generated automatically.

📋 Comprehensive Analysis Schema

The agent follows an expert-vetted reporting structure:

  1. Image Type & Region
  2. Key Findings (systematic observation)
  3. Diagnostic Assessment (evidence-backed)
  4. Patient-Friendly Explanation
  5. Research Context (PubMed-adjacent grounding)

Important

Disclaimer: This tool is for educational and informational purposes only. All AI-generated analyses must be reviewed and validated by a qualified medical professional.


AI-Assisted Diagnostics for the Modern Radiologist