Nonlinear time series analysis for on-line handwritten signature verification

In this article, an approach for online signature verification for person authentication from nonlinear time series analysis based on delay coordinate embedding has been presented. The proposed approach is found to be more effective compared to the most popular DTW based approach from the simulation experiments with the benchmark data set SVC 2004 of signature verification contest held in 2004. The proposed approach is also shown to be specially effective when tuned to individual writer which seems to be important in developing person specific algorithms in personalized authentication applications. Moreover the approach can also be extended to other behavioural biometric as such human gait.