jaekookang / attention-monitor

Our solution is a computer vision system that translates facial behavior statistics into an engagement metrics as a feedback to inform a teacher about students’ engagement in an online class.

Online Classroom Attention Tracker (OCAT)

Client: python main.py --userid 123

python main.py --userid kenghooi --host 10.10.0.106

• Real time data visualization with PyQtGraph
• Face landmarks -eye, mouth monitor
• Head pose estimation
• Server-side, Client-side data streaming(pub,sub)
• Post-session Attention Metrics Dashboard
• face,person detection - attendance
• human pose estimation - engagement, ergonomics

Contributors

William Ardianto, Kenghooi Teoh, Leonard Loh, Choo Wilson.

Pretrained Weights

• Dlib face landmarks weight - GDRIVE
• Dlib face landmarks weight
• Hopenet pretrained weight

References

• Boilerplate opencv python code, read frame from webcam, image, video. Show me the code
• Hopenet by Nataniel Ruiz
• PyImageSearch Facial landmarks with dlib, OpenCV, and Python
• PyImageSearch Drowsiness detection with OpenCV