WakeApp is a system that can automatically detect a driver drowsiness in a real-time video stream, using computer vision and neural network (deep learning) algorithms.

The classification is based on blinks, yawns, current time, and travel duration. When the driver appears to be drowsy, the system will play an alarm in the car and send a warning email to an emergency contact.

The main folder contains Data and Scripts directories, a README file, and an example video.

Data folder includes the alarm sound, email warning message, shape predictor, and background and logo images. Dataset contains classified images (train, validation and test), and Model includes summary and output, confusion matrix and scores, loss and accuracy plots, and the yawn detection model itself.

Scripts folder includes the following Python scripts:

The first GUI page is start_page.py, which reads the user details and starts the program. The second page is drowsiness_classification.py, which detects the driver drowsiness and displays the video frames.

The blink and yawn detection functions are in blink_score.py and yawn_score.py respectively. The counter thresholds are calculated in thresholds.py. The sound and email functions are listed in drowsiness_alert.py.

The yawn classification model was built with convolutional_neural_network_model.py, and its scores were evaluated using convolutional_neural_network_metrics.py.

To start the app, run start_page.py.