Online signature verification using signature down-sampling and signer-dependent sampling frequency

Abstract Online signature verification considers signatures as time sequences of different measurements the signing instrument. These signals are captured on digital devices and therefore consist a discrete number samples. To enrich or simplify this information, several verifiers employ resampling interpolation preprocessing step to improve their results; however, design decisions may be difficult generalize. This study investigates direct effect sampling rate input accuracy online systems without using techniques proposes novel system based signer-dependent frequency. Twenty verifier configurations were created for five public databases variety popular approaches evaluated 20–40 rates. Our results show that there is an optimal range frequency sample points minimizes error verifier. A 15–50 Hz point count 60–240 provided best accuracies in our experiments. As expected, lower ranges showed inaccurate interestingly, higher frequencies often decreased accuracy. The one can achieve better at least same faster by down-sampling before further processing. proposed achieved competitive state-of-the-art We also studied choosing individual each signer was tested 500 methods improved 92% test cases compared usage original