International Journal of Scientific Research and Engineering Development

As user authentication moves beyond passwords, gesture-based biometrics have emerged as a promising alternative. This paper presents GestureAuth, a camera-only hand-gesture authentication system that uses Media Pipe for hand landmark tracking and OpenCV for live gesture verification. The system allows users to register a unique hand gesture and authenticate by performing it in front of a webcam. It extracts 21 hand landmarks, computes distance and angle features, and compares them with encrypted templates stored during registration. Authentication succeeds when the live gesture matches the stored template above a defined similarity threshold. The system supports both one-hand and two-hand gesture modes, with the latter providing stronger security. The results show that the proposed approach offers a secure, contactless, and password-free authentication method that works effectively under different lighting and positioning conditions.