Localization of signatures in continuous Air writing

Data and information security plays an important role in today's IT-enabled societies. People are becoming more conversant with e-services, including digital payment and online banking. Especially, in the context of smart city or smart village, such services are becoming highly popular in countries like India. However, conventional signature still remains the most preferred choice for banking related transactions due to its robustness, though, slowly people have started adopting technologies in such services. Online air signature is one such way of introducing smart interface to e-services. This has received significant attention of the research community due to the emergence of low-cost depth sensors. They can be used for implementation of touch-less biometric authentication systems in 3D. Such systems do not require keys or passwords in the systems to prove the identities. Moreover, such systems are resilient against the stolen passwords or loss of passwords. In this paper, we propose a Leap motion sensor guided online 3D signature analysis system that is robust in nature by allowing a user to perform random gestures before and/or after the signature during authentication. Signatures can appear within any position of long gesture patterns. However, it is important to correctly spot the actual signatures for authentication. We have proposed a signature spotting mechanism using a window-based analysis on high-level features extracted from raw signatures. An efficient searching strategy has been proposed using 3D convex hull points. Dynamic Time Warping (DTW) and Hidden Markov Model (HMM) have been used to perform the verification of the spotted signatures. It has been observed that the proposed method works with more than 85% accuracy in signature spotting with less computational burden.