A Novel Sigma-Lognormal Parameter Extractor for Online Signatures

Online signature analysis can be widely applied in e-security and health. The latest method combines the Sigma-Lognormal model and visual feedback to extract the kinematic and spatial parameters of online signatures, but the model still does not perform well in complex handwriting signatures. Inaccurate parameters cannot reveal health information about users and cannot correctly reconstruct the online signature. This paper presents a novel Sigma-Lognormal parameter extractor for this drawback. On the one hand, this extractor estimates the parameters of pen-up and optimizes the parameters without the stroke midpoint. On the other hand, the extractor dynamically corrects the salient point position deviation by the velocity minimum point and velocity intersection point of adjacent strokes. The new extractor solves the parameter distortion caused by the ignored pen-ups and the hidden time deviation. The experiments demonstrate the accuracy and robustness of our method on multiple databases and verifiers, and the results show that the performance of the new extractor is better than the state-of-the-art method.