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.
- 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.
- Model: Gemini 2.0 Flash (Vision-capable)
- Framework: Agno
- Search Tool: DuckDuckGo Tools
- Image Processing: PIL (Pillow)
- UI: Streamlit
- Configure API Key:
Provide your
GOOGLE_API_KEYin the environment. - Launch:
streamlit run medical.py
- Analyze Image:
- Upload a scan (e.g., a chest X-ray).
- Clinical analysis will be generated automatically.
The agent follows an expert-vetted reporting structure:
- Image Type & Region
- Key Findings (systematic observation)
- Diagnostic Assessment (evidence-backed)
- Patient-Friendly Explanation
- 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