Fast signature spotting in continuous air writing

Abstract

Development of depth sensors paves way for implementation of touch-less biometric authentication systems using 3D gestures or signatures. No keys or passwords are required in such systems to prove identities. Moreover, the systems do not suffer from various security risks such as stolen passwords or loss of passwords. In this paper, we propose such a security system that is robust in nature by allowing a user to perform random gestures before and after a signature during authentication. Since the signature can appear within any position of a gesture pattern, correct spotting of the actual signature is extremely important. In this paper, we propose a signature spotting mechanism that has been accomplished using a window-based analysis on feature sequence. An efficient searching strategy has been proposed using 3D convex hull points. Dynamic Time Warping (DTW) has been used to perform the verification of the spotted signatures. Our proposed method achieves 80% accuracy for signature spotting with less computational overhead. The method can be used, on applications requiring robust authentication in a huge dataset. © 2017 MVA Organization All Rights Reserved.