Online signature verification using segmented local features
Madiha Tahir, Muhammad Usman Akram, Muhammad Arslan Idris · 2016
Biometric is a reliable system that provides person's authentication and verification using his physical and behavioral characteristics. Most of the previous work has been done on online signature verification involved statistical local and global features extracted from x, y coordinates and pen's pressure. Pen's positions along the time series has not used for verification yet. This motivates us to design an efficient online signature verification system by analyzing Local features consists of x, y coordinates and pen's ups and downs. Features are selected using segmentation such that characteristics of the shape of the signature remain maintained. Euclidean distance and dynamic time warping (DTW) are used for enrollment and matching of questioned sample with reference samples. Proposed method is evaluated on Japanese Online test set from ICDAR2013. System's performance is evaluated by computing EER, FAR, FRR, accuracy and time in seconds. We achieved 78.14% accuracy with average elapsed time of 0.13s using segmentation and 78.57% without segmentation with average elapsed time of 3.42s.