Alerta- Application to detect fatigue among vehicle drivers

How could we use deep learning to recognize fatigue signs among drivers in order to prevent vehicle crashes?In 2020, almost 4,014 died in large fatal truck crashes. The main goal of this project is to replace traditional ways of detecting fatigue, a primary cause of vehicle crashes, by a modern solution based on AI. While methods such as measuring distance of eyelid and eye tracking have proven their inefficiency, Computer Vision(C.V) seems to hold a great promise in achieving better results. Artificial Intelligence could be defined as a simulation of how the human learning process works. It refers to the ability of a machine to executes certain tasks without being explicitly programmed to do them.

The focus on detecting fatigue signs and not other factors in vehicle crashes is mainly because of the high correlation between the possibility of a road accident and the driver’s consciousness state. In the majority of cases, driving while feeling tired or dizzy is associated with a fatal crash risk.Since fatigue is a major factor contributing to vehicle crashes, we need to develop a solution that is able to automate the task of detecting fatigue signs among vehicle drivers. As we are recognizing signs on given images, we need to use deep learning and precisely train a CNN to perform the facial expression detection. However, the CNN should be accurate enough to deliver satisfying results and this could be assessed through different evaluation metrics