3D Pattern Authentication for Fingertip input with Leap Motion

• Main Authors: Shen-Yen Huang, 黃聖元
• Other Authors: Mau-Tsuen Yang
• Format: Others
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/12225298190863323357

碩士 === 國立東華大學 === 資訊工程學系 === 103 === Nowadays, traditional passwords, handwritten signatures, and biometric features are widely used for personal authentications. The traditional passwords are easy to implement but can suffer from random cheat attacks. The handwritten signatures are popular but fragile under observed cheat attacks. Alternatively, biometric features such as iris and fingerprint are nonintrusive but at the risk of losing users’ biometric identities. Therefore, we proposed a 3D signature recognition system using a Leap Motion to detect and track 3D fingertip positions of users. By analyzing the input 3D signature, we extracted night kinds of online features: velocities, accelerations, turning angles, coordinates and direction vectors of three axis, and a offline feature which calculate the difference by two signatures Hausdorff distance. Finally, the proposed system improved the accuracy of signature verification by combining the online features and offline features with adaptive individual weights.