Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

project - ai medical imaging diagnosis agent #55

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
66 changes: 66 additions & 0 deletions ai_agent_tutorials/ai_medical_imaging_agent/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,66 @@
# Medical Imaging Diagnosis Agent

A Medical Imaging Diagnosis Agent build on phidata powered by Gemini 2.0 Flash Experimental that provides AI-assisted analysis of medical images of various scans. The agent acts as a medical imaging diagnosis expert to analyze various types of medical images and videos, providing detailed diagnostic insights and explanations.

## Features

- **Comprehensive Image Analysis**
- Image Type Identification (X-ray, MRI, CT scan, ultrasound)
- Anatomical Region Detection
- Key Findings and Observations
- Potential Abnormalities Detection
- Image Quality Assessment
- Research and Reference

## How to Run

1. **Setup Environment**
```bash
# Clone the repository
git clone https://github.com/Shubhamsaboo/awesome-llm-apps.git
cd ai_agent_tutorials/ai_medical_imaging_agent

# Install dependencies
pip install -r requirements.txt
```

2. **Configure API Keys**
- Get Google API key from [Google AI Studio](https://aistudio.google.com)

3. **Run the Application**
```bash
streamlit run ai_medical_imaging.py
```

## Analysis Components

- **Image Type and Region**
- Identifies imaging modality
- Specifies anatomical region

- **Key Findings**
- Systematic listing of observations
- Detailed appearance descriptions
- Abnormality highlighting

- **Diagnostic Assessment**
- Potential diagnoses ranking
- Differential diagnoses
- Severity assessment

- **Patient-Friendly Explanations**
- Simplified terminology
- Detailed first-principles explanations
- Visual reference points

## Notes

- Uses Gemini 2.0 Flash for analysis
- Requires stable internet connection
- API usage costs apply
- For educational and development purposes only
- Not a replacement for professional medical diagnosis

## Disclaimer

This tool is for educational and informational purposes only. All analyses should be reviewed by qualified healthcare professionals. Do not make medical decisions based solely on this analysis.
157 changes: 157 additions & 0 deletions ai_agent_tutorials/ai_medical_imaging_agent/ai_medical_imaging.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,157 @@
import os
from PIL import Image
from phi.agent import Agent
from phi.model.google import Gemini
import streamlit as st
from phi.tools.duckduckgo import DuckDuckGo

if "GOOGLE_API_KEY" not in st.session_state:
st.session_state.GOOGLE_API_KEY = None

with st.sidebar:
st.title("ℹ️ Configuration")

if not st.session_state.GOOGLE_API_KEY:
api_key = st.text_input(
"Enter your Google API Key:",
type="password"
)
st.caption(
"Get your API key from [Google AI Studio]"
"(https://aistudio.google.com/apikey) 🔑"
)
if api_key:
st.session_state.GOOGLE_API_KEY = api_key
st.success("API Key saved!")
st.rerun()
else:
st.success("API Key is configured")
if st.button("🔄 Reset API Key"):
st.session_state.GOOGLE_API_KEY = None
st.rerun()

st.info(
"This tool provides AI-powered analysis of medical imaging data using "
"advanced computer vision and radiological expertise."
)
st.warning(
"⚠DISCLAIMER: This tool is for educational and informational purposes only. "
"All analyses should be reviewed by qualified healthcare professionals. "
"Do not make medical decisions based solely on this analysis."
)

medical_agent = Agent(
model=Gemini(
api_key=st.session_state.GOOGLE_API_KEY,
id="gemini-2.0-flash-exp"
),
tools=[DuckDuckGo()],
markdown=True
) if st.session_state.GOOGLE_API_KEY else None

if not medical_agent:
st.warning("Please configure your API key in the sidebar to continue")

# Medical Analysis Query
query = """
You are a highly skilled medical imaging expert with extensive knowledge in radiology and diagnostic imaging. Analyze the patient's medical image and structure your response as follows:

### 1. Image Type & Region
- Specify imaging modality (X-ray/MRI/CT/Ultrasound/etc.)
- Identify the patient's anatomical region and positioning
- Comment on image quality and technical adequacy

### 2. Key Findings
- List primary observations systematically
- Note any abnormalities in the patient's imaging with precise descriptions
- Include measurements and densities where relevant
- Describe location, size, shape, and characteristics
- Rate severity: Normal/Mild/Moderate/Severe

### 3. Diagnostic Assessment
- Provide primary diagnosis with confidence level
- List differential diagnoses in order of likelihood
- Support each diagnosis with observed evidence from the patient's imaging
- Note any critical or urgent findings

### 4. Patient-Friendly Explanation
- Explain the findings in simple, clear language that the patient can understand
- Avoid medical jargon or provide clear definitions
- Include visual analogies if helpful
- Address common patient concerns related to these findings

### 5. Research Context
IMPORTANT: Use the DuckDuckGo search tool to:
- Find recent medical literature about similar cases
- Search for standard treatment protocols
- Provide a list of relevant medical links of them too
- Research any relevant technological advances
- Include 2-3 key references to support your analysis

Format your response using clear markdown headers and bullet points. Be concise yet thorough.
"""

st.title("🏥 Medical Imaging Diagnosis Agent")
st.write("Upload a medical image for professional analysis")

# Create containers for better organization
upload_container = st.container()
image_container = st.container()
analysis_container = st.container()

with upload_container:
uploaded_file = st.file_uploader(
"Upload Medical Image",
type=["jpg", "jpeg", "png", "dicom"],
help="Supported formats: JPG, JPEG, PNG, DICOM"
)

if uploaded_file is not None:
with image_container:
# Center the image using columns
col1, col2, col3 = st.columns([1, 2, 1])
with col2:
image = Image.open(uploaded_file)
# Calculate aspect ratio for resizing
width, height = image.size
aspect_ratio = width / height
new_width = 500
new_height = int(new_width / aspect_ratio)
resized_image = image.resize((new_width, new_height))

st.image(
resized_image,
caption="Uploaded Medical Image",
use_container_width=True
)

analyze_button = st.button(
"🔍 Analyze Image",
type="primary",
use_container_width=True
)

with analysis_container:
if analyze_button:
image_path = "temp_medical_image.png"
with open(image_path, "wb") as f:
f.write(uploaded_file.getbuffer())

with st.spinner("🔄 Analyzing image... Please wait."):
try:
response = medical_agent.run(query, images=[image_path])
st.markdown("### 📋 Analysis Results")
st.markdown("---")
st.markdown(response.content)
st.markdown("---")
st.caption(
"Note: This analysis is generated by AI and should be reviewed by "
"a qualified healthcare professional."
)
except Exception as e:
st.error(f"Analysis error: {e}")
finally:
if os.path.exists(image_path):
os.remove(image_path)
else:
st.info("👆 Please upload a medical image to begin analysis")
5 changes: 5 additions & 0 deletions ai_agent_tutorials/ai_medical_imaging_agent/requirements.txt
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
streamlit==1.40.2
phidata==2.7.3
Pillow==10.0.0
duckduckgo-search==6.4.1
google-generativeai==0.8.3