How soummo19/facialexpressiondetector works
This is a Flask web app that streams webcam frames annotated with one of seven facial-expression labels. A browser loads the page and requests the video feed; the server captures frames, detects faces, prepares each face for the classifier, annotates the frame, and streams it back. The model and Haar-cascade files supply runtime data. The README describes the app’s purpose, while the sampled Python files show the request and inference path; the template contents were not provided, so its behavior is limited to the documented render call.
Main entry points: camera.py Video camera · main.py Flask routes · model.py Expression model · index.html Web page
Components 12
Web streaming
- Flask routes
main.py - Web page
templates/index.html - Frame generator
main.py
Vision pipeline
- Video camera
camera.py - Face detection
camera.py - Haar cascade
haarcascade_frontalface_default.xml - Frame processing
camera.py
Expression classifier
- Expression model
model.py - Model definition
model.json - Model weights
model_weights.h5
Other components
- Browser
- Webcam
Connections 13
- Browser → Flask routes: requests page/feed
README.md - Flask routes → Web page: renders
main.py - Flask routes → Frame generator: streams frames
main.py - Flask routes → Video camera: creates
main.py - Frame generator → Video camera: requests frames
main.py - Video camera → Webcam: captures frames
camera.py - Video camera → Face detection: detects faces
camera.py - Face detection → Haar cascade: uses classifier
camera.py - Video camera → Frame processing: prepares and encodes
camera.py - Video camera → Expression model: predicts emotion
camera.py - Expression model → Model definition: loads definition
model.py - Expression model → Model weights: loads weights
model.py - Frame processing → Browser: returns JPEG stream