Authentication
Each user has an account, a protected classifier page, and a private prediction history.
Flask + CNN deployment
Upload a traffic sign image, classify it with traffic_classifier.h5, and keep every prediction linked to the authenticated user.
Register or login before using the model.
Upload a sign image and receive the predicted class.
Mark predictions as true or false for later analysis.
Review your dashboard and history inside the container database.
Each user has an account, a protected classifier page, and a private prediction history.
The app loads traffic_classifier.h5 from the project root during startup.
True/false labels are saved with predictions so the dashboard can show reviewed accuracy.