Online Signature Recognition: A Biologically Inspired Feature Vector Splitting Approach

This research introduces an innovative approach to explore the cognitive andbiologically inspired underpinnings of feature vector splitting for analyzingthe significance of different attributes in e-security biometric signaturerecognition applications. Departing from traditional methods of concatenatingfeatures into an extended set, we employ multiple splitting strategies,aligning with cognitive principles, to preserve control over the relativeimportance of each feature subset. Our methodology is applied to three diversedatabases (MCYT100, MCYT300,and SVC) using two classifiers (vector quantizationand dynamic time warping with one and five training samples). Experimentationdemonstrates that the fusion of pressure data with spatial coordinates (x andy) consistently enhances performance. However, the inclusion of pen-tip anglesin the same feature set yields mixed results, with performance improvementsobserved in select cases. This work delves into the cognitive aspects offeature fusion,shedding light on the cognitive relevance of feature vectorsplitting in e-security biometric applications.