Base on Long Short-term Memory Network for Fatigue Detection

Abstract

This paper focuses on a real-time fatigue detection flow. The system will be doing this all inside of python and building it up step by step to be able to detect a bunch of different poses and specifically signs of drowsiness. In order to do that we use a few key models and using media pipe holistic to be able to extract key points. This is going to allow us to extract key points from our face. The system uses tensorflow and keras and builds up a long short-term memory(LSTM) model to be able to predict the action which is being shown on the screen. We need to do is collect a bunch of data on all of our different key points, so we collect data on our face and save those as numpy arrays. The face detection method is based on a deep neural network using LSTM layers to go on ahead and predict that temporal component, which be able to predict action from a number of frames not just a single frame. Integrate using opencv and then proceed to make real-time predictions using the webcam.