How abdullah374/dentxai works
DentXAI’s implemented workflow is a small Streamlit dental-image detector: a user uploads an image, the app runs YOLO inference with a confidence threshold, displays annotated output, saves it, and offers a download. The README describes additional classification, explainability, batch, and evaluation capabilities, but the supplied tree and source excerpt do not show their implementation, so they are not mapped here.
Main entry points: app.py Streamlit app · outputs Annotated output files · best.pt Trained weights · uploads Uploaded images
Components 7
Other components
- Dental professional
- Streamlit app
app.py - Trained weights
runs/detect/intraoral/weights/best.pt - Ultralytics YOLO
- Uploaded images
uploads - PIL image conversion
- Annotated output files
outputs
Connections 11
- Dental professional → Streamlit app: uploads image
app.py - Streamlit app → Trained weights: loads weights
app.py - Streamlit app → Ultralytics YOLO: creates model
app.py - Streamlit app → Uploaded images: writes upload
app.py - Streamlit app → Ultralytics YOLO: runs prediction
app.py - Ultralytics YOLO → Streamlit app: returns detections
app.py - Streamlit app → Ultralytics YOLO: plots detections
app.py - Streamlit app → PIL image conversion: converts image
app.py - Streamlit app → Annotated output files: saves result
app.py - Streamlit app → Annotated output files: reads result
app.py - Streamlit app → Dental professional: shows and offers download
app.py