We chose to make this project because we considered the road accident that took place in the night.
This project detects the eye shape of the person in front of the camera, which we might assume to be a driver, and detects if the person is sleeping or not to avoid accidents in night.
This project is a machine learning project that is built upon image recognition algorithms. we used a few Python libraries, one of them is 'dlib' which is a library that helps predict the eye shape from a '.dat' file which is the dataset used for this machine learning model to train. We used 'imutils' and 'face_utils' to clearly understand the shape of the eyes,' pygamer' library's mixer module to play the alarm when the code detects that the person is asleep, and the last and the main library used in the project is 'opencv' which is an open-source python library which acts as a computer vision.
There was one main challenge that we faced while making the project which was regarding to the database. While researching for the datasets we came to know that most of the datasets around the internet were of jpg format and they contained thousands of eye shapes in jpg format. We found that the data preprocessing for this type of dataset and it took a lot of time for us to integrate the preprocessing of data.
The accomplishment that we are proud of about the project is that the code accurately detects the person's sleep face and awake face.
we learned about using the datasets properly.
for the plans for this project in the future, we are planning on making it a fully accessible Android app. While the driver places the mobile on the car's dashboard, or the tripod stand the app runs in the background and alarms the driver when they are asleep.
Log in or sign up for Devpost to join the conversation.