Drowsiness has been involved in about 10-40% of the accidents on roads. Fall-asleep accidents are very serious in terms of the severity of the injury and is more likely to occur in people who are sleep deprived. Drowsiness affects a person’s attentiveness, also it reduces their capability to handle a vehicle safely. In addition, it associates to slow response time, decreased awareness and imprecise judgment. Signs of drowsiness such as tiredness in eyes and yawning are a clear indication of drivers losing control of themselves. We can use these signs to alert them timely, so that they can immediately regain their control. This can help prevent accidents. We propose a system that computes EAR(Eye Aspect Ratio) and MAR(Mouth Aspect Ratio) to detect drowsiness. We extract the key-points of eyes and mouth using Mediapipe to calculate EAR and MAR.
