Face Img/Video based drowsiness recognition

This repository is dedicated to “AUTOMOTIVE” Project. This was a collaboration with Visual Computing and Machine Intelligence (VCMI) Research Group of | INESC Technology and Science | Porto.

Goal

Development of robust real-time Drowsiness detector system to detect driver fatigue:

• Explore impact of facial dynamic fusion information (landmarks and eyes/mouth images)
• Compare the use of DBN with traditional ML classifier (SVM, KNN, DecisionTree)

Approach

Development ML\DL models for detecting drowsiness recognition based on facial image/video.

Scripts/Files

• pp_presentation (pdf/slides that summarize the internship results and were used to support the oral presentation)
• python_notebook (python implementation: preprocessing, frames visualization, Training loop and testing)