How bavan1223/convnext_oral_cancer_detection works
This Flask application accepts oral-lesion images, screens and classifies them, then presents a clinical-style result. The README also documents Grad-CAM heatmaps, but the sampled source does not show their implementation or wiring; that documented capability is shown separately. The model weights and upload directory are not present in the tree, so no separate repository paths are assigned to them.
Main entry points: app.py Flask routes · index.html Upload page · script.js File check · result.html Result page
Components 14
Web interface
- Upload page
templates/index.html - File check
static/js/script.js - Flask routes
app.py - About page
templates/about.html - Contact page
templates/contact.html
Image analysis
- Uploaded images
app.py - Oral image filter
app.py - Image preprocessing
app.py - ConvNeXt classifier
app.py
Diagnosis and review
- Clinical classification
app.py - Result page
templates/result.html - Grad-CAM heatmaps
Other components
- Image submitter
- Clinician reviewer
Connections 12
- Image submitter → Upload page: opens
- Upload page → File check: checks file
- Image submitter → Flask routes: submits image
- Flask routes → Uploaded images: saves image
- Flask routes → Oral image filter: checks image
- Oral image filter → Flask routes: returns validity
- Flask routes → Image preprocessing: transforms image
- Flask routes → ConvNeXt classifier: runs inference
- Flask routes → Clinical classification: assigns diagnosis
- Flask routes → Result page: renders result
- ConvNeXt classifier → Grad-CAM heatmaps: generates heatmap
- Grad-CAM heatmaps → Clinician reviewer: supports review