Online Signature Verification Using Locally Weighted Dynamic Time Warping via Multiple Fusion Strategies

Online signature verification is the process of using a dynamic system to confirm writer’s identity. It can be used as security entrance applications and password substitutes, well forensic tool assist an expert’s investigation. This study proposes novel online based on single-template strategy improve performance in real-world scenarios. employs discriminative mean template sets fusion strategies multiple local weighting warping schemes, for time (DTW). The first step generate set user-specific templates each feature recent time-series averaging method, namely, Euclidean barycenter-based DTW barycenter averaging. Then, direct matching points between references dependent independent DTW, we obtain estimate considering stability sequences. Furthermore, develop calculating locally weighted concatenating them vector warping, followed by construction support machine (SVM) classifier. Finally, phase, use technique compute fused score SVMs query sample. effectiveness proposed method demonstrated extensive experimental results obtained three public datasets: SVC2004 Task1/Task2 MCYT-100.