.png)
.png)
Email this article 
Pre-processing of handwritten signature images for following recognition
https://doi.org/10.21122/2309-4923-2022-2-4-9
Abstract
In the process of handwritten signature recognition, preprocessing is an important step before calculation its informative features. Signatures of any person always have some differences, in addition, they can be of different colors, sizes and orientations. After signature digitization, their images may contain some noise. The purpose of signature image preprocessing is to obtain the most invariant representation of the digital image of a person's signature, which will allow us to identify him or define that the signature is forged.
This paper describes a sequence of transformations necessary to perform preprocessing of the signature image and form its representation of a single orientation and size. It is assumed that there are no graphic elements in the image that are not related to the signature and the background is relatively uniform. The transformations under consideration sequentially perform binarization of the signature image, its filtering, rotation, cropping of the circumscribed rectangle and scaling to a fixed size.
The described preprocessing procedures were applied to a number of available digital signature image databases, such as CEDAR, BHSig260-Bengali, BHSig260-Hindi. Experiments on signature recognition confirm that the presented approach to the signature image preprocessing leads to increasing of the recognition accuracy.
About the Authors
U. Yu. AkhundjanovBelarus
Akhundjanov Umidjon Yunus ugli, PhD student
V. V. Starovoitov
Belarus
Starovoitov Valery, Doctor of technical sciences, professor
References
1. Hafemann, L.G. Offline handwritten signature verification — Literature review / L.G. Hafemann, R. Sabourin, L.S. Oliveira // Seventh International Conference on Image Processing Theory, Tools and Applications (IPTA) – 2017. 8 p. DOI:10.1109/ipta.2017.8310112.
2. Plamondon, R. Online and off-line handwriting recognition / R. Plamondon, S. Srihari // IEEE Transactions on Pattern Analysis and Machine Intelligence, – 2000. – Vol. 22, № 1. – P. 63–84.
3. Ferrer, M. A. A behavioral handwriting model for static and dynamic signature synthesis / M. A. Ferrer, M. Diaz, C. Carmona-Duarte,cA. Morales // IEEE Transactions on Pattern Analysis and Machine Intelligence, – 2017. – Vol. 39, № 6. – P. 1041–1053.
4. Gonsales, R. Cifrovaja obrabotka izzobrajeniy / R. Gonsales, R. Vuds. – М.: Tehnosfera, 2005. – 1072 p.
5. Shapiro, L. Komp’juternoe zrenie / L.Shapiro, Dzh.Stokman; Per. s angl. — 3-е izd. — M.: BINOM. Labotatotija znaniy, 2015. – 763 s.
6. Gruzman, I.S. Cifrovaja obrabotka izobrazheniy v informacionnyh: ucheb. posobie / I.S. Gruzman, V.S. Krichuk, V.P. Kosyh, G.I. Peritjagin, A.А. Spektor. – Novosibirsk.: Novosib. gos. tehn. un-t., 2003. – 352 s.
7. Coetzer, J. Off-line signature verification using the discrete radon transform and a hidden Markov model / J. Coetzer, B.Herbst, J. Preez // EURASIP Journal on Advances in Signal Processing, – 2004. – Vol. 2004, № 4. – P. 559–571.
8. Pallavi, V. Hatkar. Image Processing for Signature Verification / V. Hatkar Pallavi, Zareen J Tamboli // International Journal of Innovative Research in Computer Science & Technology (IJIRCST), – 2015. – Vol. 3, № 3. – P. 127–129.
9. Goritov, A.N. Preliminary image processing in technical vision systems / A.N. Goritov // Doclady TUSUR. – 2018. – Vol. 21, № 4-1. – P. 53-58
10. Dataset [Electronyj resurs]. – Rezhim dostupa: https://drive.google.com/file/d/0B29vNACcjvzVc1RfVkg5dUh2b1E/view. Data dostupa: 28.04.2022
11. cedar.buffalo.edu [Electronyj resurs]. – Rezhim dostupa: http://www.cedar.buffalo.edu/NIJ/data/signatures.rar. Data dostupa: 28.04.2022
12. Recommendation ITU-R BT.709-6. Parameter values for the HDTV standards for production and international programme exchange. – 2015. – P. 19.
13. Otsu, N. A threshold selection method from gray-level histograms// IEEE Transactions on Systems, Man and Cybernetics. – 1979. – Vol. 9, № 1. – P. 62–66.
14. Gasmi, A. PCA and SVM as geo-computational methods for geological mapping in the southern of Tunisia, using ASTER remote sensing data set / A. Gasmi, C. Gomez, H. Zouari, A. Masse, D. Ducrot // Arabian Journal of Geosciences, – 2016. – Vol. 9, № 20. – Num. 753. https://doi.org/10.1007/s12517-016-2791-1.
15. Du, Q. Hyperspectral image compression using JPEG2000 and principal component analysis / Q. Du, J. E. Fowler // IEEE Geoscience and Remote Sensing Letters, – 2007. – Vol. 4, № 2. – P. 201–205.
16. Zhu, Y. N. Improved principal component analysis and linear regression classification for face recognition / Y. N. Zhu, C. Y. Zhu, X. Li // Signal Process., – 2018. – Vol. 145. – P. 175–182.
17. Lai, Z. Multilinear sparse principal component analysis / Z. Lai, Y. Xu, Q. Chen, J. Yang, D. Zhang // IEEE Transactions on Neural Networks and Learning Systems, – 2014. – Vol. 25, No. 10. – P. 1942–1950.
18. Yi, S. Joint sparse principal component analysis / S. Yi, Z. Lai, Z. He, Y.-M. Cheung, Y. Liu // Pattern Recognition, – 2017. – Vol. 61. –P. 524–536.
19. Artoni, F. Applying dimension reduction to eeg data by principal component analysis reduces the quality of its subsequent independent component decomposition / F. Artoni, A. Delorme, S. Makeig // NeuroImage, – 2018. – Vol. 175. –P. 176–187.
Supplementary files
For citation: Akhundjanov U.Y., Starovoitov V.V. Pre-processing of handwritten signature images for following recognition. «System analysis and applied information science». 2022;(2):4-9. https://doi.org/10.21122/2309-4923-2022-2-4-9
Refbacks
- There are currently no refbacks.
