Due to the recent outbreak of COVID-19, online courses are gaining popularity these days. But they aren’t as effective as classroom teaching. Students may fall asleep during the classes. Even the employees who work from home tend to feel drowsy during work. Using a vision-based technique is one of the technical options for implementing drowsiness detection systems. Here, we are detecting the person’s drowsiness by estimating the vision system. The drowsiness detection using the computer vision model is designed to identify whether the student is drowsy or not. Computer vision enables us to extract the features that are required to predict drowsiness in a person, and alert the person through some signal. The alert we wanted to trigger was an alarm system, which possesses some types of puzzle or math problem which needs to be solved in order to stop the alarm sound. By solving this puzzle, the person may stop being sleepy. In this model, we extract live face images from the webcam using OpenCV; by applying “68 face landmark detector” from the Dlib library, we detect 68 unique points on the face, and by calculating the distance between the points corresponding to the eye, the model predicts the state of the person.