The drowsiness detection system described here combines machine learning techniques, Python libraries such as NumPy, SciPy, imutils, and dlib, an Arduino-based BPM sensor, and a Flask-based frontend. This system aims to monitor a person's drowsiness by analysing their facial landmarks. This system aims to monitor a person's drowsiness by analysing their facial landmarks.
The machine learning model is primarily built using the dlib library, which offers pre-trained facial landmark detection models. These models enable the extraction of facial features like eye positions, mouth movements, and head pose. By analyzing these features, the model can identify signs of drowsiness.
The model collects the facial landmark detection from dlib and utilizes NumPy and SciPy for data manipulation and analysis, enabling the extraction of relevant features. These features are then fed into the machine learning algorithm neural network, to train the drowsiness detection model.
Once the model is trained, it can be deployed using a Flask-based frontend. The frontend provides a user interface where the drowsiness detection system can be accessed and monitored. It displays real-time information about the person's state whether drowsy or awake.
The Flask framework allows seamless integration between the trained model and the frontend interface. It enables the display of real-time predictions and alerts when drowsiness levels exceed a certain threshold, ensuring timely intervention to prevent accidents caused by drowsiness while driving or operating machinery.
To collect the Heart-Rate of the person, an Arduino-based BPM sensor is used to measure the person's heart rate. The BPM sensor communicates with the Python code through a serial connection, and the BPM is calculated using
Overall, this system combines facial landmark detection, heart rate monitoring, and machine learning algorithms to create an effective drowsiness detection solution. It provides a comprehensive approach to monitoring drowsiness levels in real-time and can be further customized and enhanced to meet specific requirements or integrate with other technologies.