Today, most of us have already used bio-metric authentication methods, including fingerprints and facial recognition, on our phones and laptops. Nevertheless, these methods mostly require dedicated hardware, which is not available across all devices like desktop PCs and basic laptop models. Another problem is that the methods in use today are primarily used for single-time authentication. Hence, these methods are used only in the login process, and once you pass one of the authentication methods, the system will require no further authentication from any user on a specific device. Most computers (PCs and laptops) are not equipped with the hardware required for facial recognition and fingerprint scanning. Therefore, an alternative biometric authentication method thatdoesnotrequireanydedicatedhardwarewillbenefittheaverageuserandthemanufacturing company by eliminating the need to include these hardware additions in the product, which will lower the production costs for the company and hopefully for the company and the customer. This project will present an approach that focuses on the most common hardware used today, the keyboard. Our approach is that by analyzing keystroke gestures, we can get a sufficient authentication of a user at the login process and during his use of the app/website. The project document is arranged as follows. Section 2 discusses relevant background topics. Section 3 describes the technologies we used in our project. Section 4 will elaborate on the dataset we used and the data set we collected. Section 5 discusses the preprocessing we did on the data before starting the training. Section 6 will explain the models and their architectures. Lastly, Section 7 will describe the results of our research and will present what we achieved in comparison with other works
For more information please open the BioKey.pdf file.