User Authentication and Veri cation API with Facial and Gesture Recognition

Abstract

Using passwords, pin-codes, facial recognition, voice recognition, and the OTP (One Time Password) two way veri cation protocol, work ne but are subject to easy access due to human errors. Thus, it is very easy for a hacker to gain access to systems using social engineering techniques. According to SANs, the weakest vulnerability in every system is the end user. We propose a system to patch and reduce the human errors that make these user authentication and veri cation systems vulnerable. A proposed security API system, veri es user identity remotely and instantaneously using arti cial intelligence and facial recognition authentication process where a real-time image and video feed of the user’s face will be matched by comparing the records of the user in a software’s database (three-way veri cation). This is different from the usual facial recognition system as it uses real-time facial gestures by using different levels of security during user authentication and veri cation. The standard password authentication is an auxiliary to this new type of authentication, where the two-factor authentication is taken into consideration. The real-time face capture does not take pictures but rather records the facial gestures of the end-user. This is done to prevent hackers from using images to trick the system into thinking the authentication is satis ed. According to research, it is realized that people give out their passwords and pin-codes to others to perform transactions on their behalf. For example, in the case of banking, a user can give out his or her banking credentials to his or her friend to make a withdrawal on his or her behalf. In this scenario, there is no way the banking system will be able to identify the person making the withdrawal since it assumes it is the actual customer due to a successful login. The propose system eliminates these aws in existing authentication systems by adding a visual authentication.