Drowsiness is one of the major causes of traffic accidents. This paper presents a simple and effective drowsiness detection system using IoT by locating the eyes from driving images and detecting drowsiness based on the eye aspect ratio obtained from the eye landmarks extracted. We conduct experiments under a variety of circumstances, such as adjusting the distance between the driver and the camera, wearing different accessories, and varying the driver’s head movement. Our system can detect eye landmarks well even when the driver covers most of his face with a mask and glasses or a hat and glasses.