SVSV: Online handwritten signature verification based on sound and vibration

Zhixiang Wei, Song Yang*, Yadong Xie, Fan Li, Bo Zhao

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)

Abstract

Handwritten signature is one of the most important behavioral biometrics and plays an important role in the field of identity verification. It is regarded as a legal means to verify personal identity by administrative and financial institutions. Traditional manual signature verification requires large labor costs and the probability of verification error is relatively high. Nowadays, tablets are often used for signature capturing, which motivates us to explore the feasibility of using tablets for signature verification. In this paper, we propose an online handwriting signature verification system based on sound and vibration (SVSV) generated during the signing process. We develop an application to collect signature-related vibration and sound data. We first extract the time domain features of the sound signal and use Fast Fourier Transform to extract the frequency domain features of the sound data. For vibration data, we use Discrete Cosine Transform for dimensionality reduction and feature extraction. Then we fuse the sound and vibration features. Finally, we design an efficient one-class classifier based on the Convolutional Neural Network to perform signature verification. Through extensive experiments with 12 volunteers, the results show that SVSV is a robust and efficient system with an AUC of 0.984 and an EER of 0.05.

Original languageEnglish
Pages (from-to)109-125
Number of pages17
JournalInformation Sciences
Volume572
DOIs
Publication statusPublished - Sept 2021

Keywords

  • Acoustic signal
  • CNN
  • Handwritten signature verification
  • Vibration

Fingerprint

Dive into the research topics of 'SVSV: Online handwritten signature verification based on sound and vibration'. Together they form a unique fingerprint.

Cite this