Signature down-sampling for finger input online signature verification

Abstract

Online signatures are widely accepted and used for authentication purposes. They are acquired using special devices with different sampling frequencies. A stylus or finger can be used as an input method. This paper studied the minimum sampling frequency required to achieve the best accuracy in online signature verification systems. Three different dynamic time warping-based verifiers were applied on the finger input signatures set of the DeepSignDB database. The results show that we can achieve a highly accurate online signature verification system for finger input signatures using lower sampling frequency, reducing time and computation costs.