Biometric Authentication Utilizing EEG Based-on a Smartphone’s 3D Touchscreen Sensor
Aseel Yasir Younis, Moceheb Lazam Shuwandy · 2023
Recently, smartphone manufacturers have begun including large storage capacities and powerful processors. Users’ personal lives, professional lives, and vast multimedia collections are all kept on their smartphones. Personal identification numbers (PINs), passwords, biometric data, gestures, and patterns are used extensively by these mobile phones. However, these mechanisms have numerous security flaws and are vulnerable to attacks, including shoulder surfing. Electroencephalography (EEG) signals are one of a kind, and this fact can be used to overcome the shortcomings of currently used methods. These signals can be captured and sent via wireless channels to be processed. In this research, we present a novel security system for mobile devices that combines the 3D touch authentication offered by 3DTPT with EEG signals. The authentication credentials based on 3D patterns are considered identification tokens. We have considered using EEG signals captured during a three-level pattern drawing over the smartphone’s screen as part of the authentication procedure. To do so, we recorded the brain activity of 10 participants as they drew one of three distinct patterns. The system’s security was tested by simulating 25 unsuccessful attempts by five malicious users to break in using the same methods as five legitimate users. A Hidden Markov Model (HMM) models EEG signals, and a Support Vector Machine (SVM) binary classifier verifies test pattern reliability. The DET, HTER, and ROC curves are security matrices used to evaluate verification systems. Positive experimental findings suggest this method has potential as an alternative to developing sophisticated authentication systems for mobile devices.