Driving Alert Application Based On Facial Expression Using Landmark Algorithm

Many factors contribute to traffic mishaps, the vast majority are the result of human error, such as fatigue. Therefore, this problem prompts the idea to monitor a driver's eyes and detect the sleepy and drowsy state of the driver. A driving alert application based on facial expression recognition is developed to materialize the idea. This application intends to help in decreasing the risks of road accidents by providing an alarm to alert users while driving in a sleepy or drowsy state. The application can determine the state of the drivers in three modes which is normal, sleepy, and drowsy by implementing the BlazeFace framework and landmark algorithm. Accuracy tests were done, and the results produced high accuracy for normal lighting conditions and moderately accurate in low light conditions.