Abstract—Handwritten signature recognition is a biometric
task used extensively in our daily life. The efficacy of such
system is important and challenging in that the recognition
accuracy still has room for improvement. In this paper, we
propose the use of Deep Convolutional Neural Networks
(DCNN), which is a deep learning technique, to improve
accuracy of handwritten signature recognition. We apply
DCNN in two difference strategies for signature recognition: 1)
transfer learning using leveraged features from a pre-trained
model on a larger dataset, and 2) create CNN model from
scratch. Our studied dataset consists of 600 pictures of
handwritten signatures collected from 30 people. In order to
evaluate the effectiveness of the proposed method, the accuracy
is compared with the results obtained from various machine
learning methods. The comparison reveals very satisfied
recognition results in the sense that the two proposed strategies
achieve 100% of the recognition rate. To compare the two
strategies in terms of training time, the strategy of creating
DCNN model from scratch shows much lower training time
than the transfer learning strategy.
Index Terms—Deep learning, deep convolutional neural
networks, handwritten signature recognition, transfer learning.
A. Hirunyawanakul, S. Bunrit, and K. Kerdprasop are with the School of
Computer Engineering, Suranaree University of Technology, Nakhon
Ratchasima 30000, Thailand (e-mail: Anusara.hi@gmail.com,
nittaya@sut.ac.th).
N. Kerdprasop is with the School of Computer Engineering and Data and
Knowledge Engineering Research Unit, Suranaree University of Technology,
Nakhon Ratchasima 30000, Thailand (e-mail: kerdpras@sut.ac.th).
[PDF]
Cite:Anusara Hirunyawanakul, Supaporn Bunrit, Nittaya Kerdprasop, and Kittisak Kerdprasop, "Deep Learning Technique for Improving the Recognition of Handwritten Signature," International Journal of Information and Electronics Engineering vol. 9, no. 4, pp. 72-78, 2019.