sapiСистемный анализ и прикладная информатика«System analysis and applied information science»2309-49232414-0481Belarusian National Technical University10.21122/2309-4923-2023-2-49-58sapi-619Research ArticleОбработка информации и принятие решенийData processing and decision–makingРаспределение значений локальной кривизны как структурный признак для off-line верификации рукописной подписиDistribution of local curvature values as a structural feature for off-line handwritten signature verificationСтаровойтовВ. В.StarovoitovV. V.

Профессор; главный научный сотрудник ОИПИ НАН Беларуси

Минск

Doctor of Sciences and professor;  Principal research fellow at the United Institute of Informatics Problems

Minsk

АхунджановУ.AkhundjanovU.

Младший научный сотрудник 

Минск

Junior research assistant

Minsk

Объединенный институт проблем информатики Национальной академии наук Беларуси; Белорусский национальный технический университетБеларусьUnited Institute of Informatics Problem of the National Academy of Sciences of Belarus; Belarusian National Technical UniversityBelarusОбъединенный институт проблем информатики Национальной академии наук БеларусиБеларусьUnited Institute of Informatics Problem of the National Academy of Sciences of BelarusBelarus202303102023024958Copyright © Старовойтов В.В., Ахунджанов У., 20232023Старовойтов В.В., Ахунджанов У.Starovoitov V.V., Akhundjanov U.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://sapi.bntu.by/jour/article/view/619

В работе предложен новый признак описания цифрового изображения рукописной подписи на базе частотного распределения значений локальной кривизны контуров этой подписи. Подробно описывается вычисление этого признака на бинарном изображении подписи. Формируется нормализованная гистограмма распределений значений локальной кривизны для 40 интервалов. Частотные значения, записанные в виде 40-мерного вектора, названы кодом локальной кривизны подписи.

При верификации близость двух подписей определяется корреляцией между кодами кривизны и LBP-кодами, описанными авторами в работе [23]. Для выполнения процедуры верификации подписи строится двумерное признаковое пространство, содержащее образы корреляционной близости пар подписей. При верификации подписи с N подлинными подписями этого же человека в признаковом пространстве представлено N(N-1)/2 образов близости пар подлинных подписей и N образов пар близости анализируемой подписи с подлинными. В качестве классификатора используется машина опорных векторов (SVM).

Экспериментальные исследования выполнены на оцифрованных изображениях подлинных и фальшивых подписей из двух баз. Точность автоматической верификации подписей на общедоступной базе CEDAR составила 99,77 %, а на базе TUIT 88,62 %.

In the paper, a new feature for describing a digital image of a handwritten signature based on the frequency distribution of the values of the local curvature of the signature contours, is proposed. The calculation of this feature on the binary image of a signature is described in detail. A normalized histogram of distributions of local curvature values for 40 bins is formed. The frequency values recorded as a 40-dimensional vector are called the local curvature code of the signature.

During verification, the proximity of signature pairs is determined by correlation between curvature codes and LBP codes described by the authors in [23]. To perform the signature verification procedure, a two-dimensional feature space is constructed containing images of the proximity of signature pairs. When verifying a signature with N authentic signatures of the same person, N(N-1)/2 patterns of the proximity of pairs of genuine signatures and N images of pairs of proximity of the analyzed signature with genuine signatures are presented in the feature space. The Support Vector Machine (SVM) is used as a classifier.

Experimental studies were carried out on digitized images of genuine and fake signatures from two databases. The accuracy of automatic verification of signatures on the publicly available CEDAR database was 99,77 % and on TUIT was 88,62 %.

кривизнабинарное контурное представление подписигистограммакорреляцияcurvaturebinary contour representation of the signaturehistogramcorrelationReferencesLin W. Y. et al. Robust and accurate curvature estimation using adaptive line integrals // EURASIP Journal on Advances in Signal Processing. – 2010. – Vol. 2010. – Pp. 1-14. DOI: 10.1155/2010/240309Lin W. Y. et al. Robust and accurate curvature estimation using adaptive line integrals. EURASIP Journal on Advances in Signal Processing, 2010, vol. 2010, pp. 1-14. DOI: 10.1155/2010/240309Cazals F., Pouget M. Estimating differential quantities using polynomial fitting of osculating jets // Computer Aided Geometric Design. – 2005. – Vol. 22, iss. 2. – Pp. 121-146. DOI: 10.1016/j.cagd.2004.09.004Cazals F., Pouget M. Estimating differential quantities using polynomial fitting of osculating jets. Computer Aided Geometric Design, 2005, vol. 22, iss. 2, pp. 121-146. DOI: 10.1016/j.cagd.2004.09.004Gao D. Integrating 3D seismic curvature and curvature gradient attributes for fracture characterization: Methodologies and interpretational implications // Geophysics. – 2013. – Vol. 78. – iss. 2. – Pp. O21-O31. DOI: 10.1190/geo2012-0190.1Gao D. Integrating 3D seismic curvature and curvature gradient attributes for fracture characterization: Methodologies and interpretational implications. Geophysics, 2013, vol. 78, iss. 2, pp. O21-O31. DOI: 10.1190/geo2012-0190.1Chen He X., Yung N.H.C. Corner detector based on global and local curvature properties // Optical engineering. – 2008. – Vol. 47, iss. 5. – Pр. 057008-057008–12. DOI: 10.1117/1.2931681Chen He X., Yung N.H.C. Corner detector based on global and local curvature properties. Optical engineering, 2008, vol. 47, iss. 5, pp. 057008-057008–12. DOI: 10.1117/1.2931681Zhang W. et al. Discrete curvature representations for noise robust image corner detection // IEEE Transactions on Image Processing. – 2019. – Vol. 28, iss. 9. – Pр. 4444-4459. DOI: 10.1109/tip.2019.2910655Zhang W. et al. Discrete curvature representations for noise robust image corner detection. IEEE Transactions on Image Processing, 2019, vol. 28, iss. 9, pp. 4444-4459. DOI: 10.1109/tip.2019.2910655Fischer P., Brox T. Image descriptors based on curvature histograms // Pattern Recognition: 36th German Conference, GCPR 2014, Münster, Germany, September 2-5, 2014, Proceedings 36. – Springer International Publishing, 2014. – Pр. 239-249. DOI: 10.1007/978-3-319-11752-2_19Fischer P., Brox T. Image descriptors based on curvature histograms. Pattern Recognition: 36th German Conference, GCPR 2014, Münster, Germany, September 2-5, 2014, Proceedings 36. Springer International Publishing, 2014, pp. 239-249. DOI: 10.1007/978-3-319-11752-2_19Садыков С.С. Алгоритм логического определения кривизны точек дискретной линии / С.С. Садыков // Алгоритмы, методы и системы обработки данных. – 2015. – №. 1. – С. 52-59.Sadykov S.S. Algorithm for the logical determination of the curvature of points of a discrete line. Algorithms, methods and data processing systems, 2015, no. 1, pp. 52-59.Worring M., Smeulders A.W.M. Digital curvature estimation // CVGIP: Image understanding. – 1993. – Vol. 58, iss. 3. – Pр. 366-382. DOI: 10.1006/ciun.1993.1048Worring M., Smeulders A.W.M. Digital curvature estimation. CVGIP: Image understanding, 1993, vol. 58, iss. 3, pp. 366-382. DOI: 10.1006/ciun.1993.1048Tong W.S., Tang C.K. Robust estimation of adaptive tensors of curvature by tensor voting //IEEE Transactions on Pattern Analysis and Machine Intelligence. – 2005. – Vol. 27, iss. 3. – Pр. 434-449.Tong W.S., Tang C.K. Robust estimation of adaptive tensors of curvature by tensor voting. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2005, vol. 27, iss. 3, pp. 434-449.Flynn P.J., Jain A.K. On reliable curvature estimation // CVPR. – 1989. – Vol. 88. – Pр. 5-9.Flynn P.J., Jain A.K. On reliable curvature estimation. CVPR, 1989, vol. 88, pp. 5-9.Trucco E., Fisher R.B. Experiments in curvature-based segmentation of range data // IEEE Transactions on Pattern Analysis and Machine Intelligence. – 1995. – Vol. 17, iss. 2. – Pр. 177-182.Trucco E., Fisher R.B. Experiments in curvature-based segmentation of range data. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1995, vol. 17, iss. 2, pp. 177-182.Magid E., Soldea O., Rivlin E. A comparison of Gaussian and mean curvature estimation methods on triangular meshes of range image data // Computer Vision and Image Understanding. – 2007. – Vol. 107, iss. 3. – Pр. 139-159. DOI: 10.1016/j.cviu.2006.09.007Magid E., Soldea O., Rivlin E. A comparison of Gaussian and mean curvature estimation methods on triangular meshes of range image data. Computer Vision and Image Understanding, 2007, vol. 107, iss. 3, pp. 139-159. DOI: 10.1016/j.cviu.2006.09.007Al-Sharadqah A., Chernov N. Error analysis for circle fitting algorithms. Electronic Journal of Statistics. – 2009. – Vol. 3. – Pр. 886-911. DOI: 10.1214/09-ejs419Al-Sharadqah A., Chernov N. Error analysis for circle fitting algorithms. Electronic Journal of Statistics, 2009, vol. 3, pp. 886-911. DOI: 10.1214/09-ejs419Gadermayr M. et al. Shape curvature histogram: A shape feature for celiac disease diagnosis // Medical Computer Vision. Large Data in Medical Imaging: Third International MICCAI Workshop, MCV 2013, Nagoya, Japan, September 26, 2013, Revised Selected Papers 3. – Springer International Publishing, 2014. – Pр. 175-184.Gadermayr M. et al. Shape curvature histogram: A shape feature for celiac disease diagnosis. Medical Computer Vision. Large Data in Medical Imaging: Third International MICCAI Workshop, MCV 2013, Nagoya, Japan, September 26, 2013, Revised Selected Papers 3. Springer International Publishing, 2014, pp. 175-184.Guerbai Y., Chibani Y., Hadjadji B. The effective use of the one-class SVM classifier for handwritten signature verification based on writer-independent parameters //Pattern Recognition. – 2015. – Vol. 48, iss. 1. – Pр. 103-113.Guerbai Y., Chibani Y., Hadjadji B. The effective use of the one-class SVM classifier for handwritten signature verification based on writer-independent parameters. Pattern Recognition, 2015, vol. 48, iss. 1, pp. 103-113.Han J. H., Poston T. Chord-to-point distance accumulation and planar curvature: a new approach to discrete curvature // Pattern Recognition Letters. – 2001. – Vol. 22, iss. 10. – Pр. 1133-1144.Han J.H., Poston T. Chord-to-point distance accumulation and planar curvature: a new approach to discrete curvature. Pattern Recognition Letters, 2001, vol. 22, iss. 10, pp. 1133-1144.Arica N., Vural F.T.Y. BAS: a perceptual shape descriptor based on the beam angle statistics // Pattern Recognition Letters. – 2003. – Vol. 24, iss. 9-10. – Pр. 1627-1639.Arica N., Vural F.T.Y. BAS: a perceptual shape descriptor based on the beam angle statistics. Pattern Recognition Letters, 2003, vol. 24, iss. 9-10, pp. 1627-1639.Coeurjolly D., Miguet S., Tougne L. Discrete curvature based on osculating circle estimation // Visual Form 2001: 4th International Workshop on Visual Form, IWVF4 Capri, Italy, May 28-30, 2001 Proceedings. – Springer Berlin Heidelberg, 2001. – Pр. 303-312.Coeurjolly D., Miguet S., Tougne L. Discrete curvature based on osculating circle estimation. Visual Form 2001: 4th International Workshop on Visual Form, IWVF4 Capri, Italy, May 28–30, 2001 Proceedings. Springer Berlin Heidelberg, 2001, pp. 303-312.Kerautret B., Lachaud J.O. Robust estimation of curvature along digital contours with global optimization // Lecture Notes in Computer Science. – 2008. – Vol. 4992. – Pр. 334-345.Kerautret B., Lachaud J.O. Robust estimation of curvature along digital contours with global optimization. Lecture Notes in Computer Science, 2008, vol. 4992, pp. 334-345.Alam S. et al. One-class support vector classifiers: A survey // Knowledge-Based Systems. – 2020. – Vol. 196. – Pр. 105754. DOI: https://doi.org/10.1016/j.knosys.2020.105754Alam S. et al. One-class support vector classifiers: A survey. Knowledge-Based Systems, 2020, vol. 196, pp. 105754. DOI: 10.1016/j.knosys.2020.105754Ахунджанов У.Ю. Предварительная обработка изображений рукописных подписей для последующего распознавания / У.Ю. Ахунджанов, В.В. Старовойтов // Системный анализ и прикладная информатика. – 2022. – № 2. – С. 4-9.Akhundjanov U.Yu., Starovoitov V.V. Pre-processing of handwritten signature images for following recognition. Sistemnyj analiz i prikladnaja informatika [System Analysis and Applied Information Science], 2022, no. 2, pp. 4-9 (In Russ.). DOI: 10.21122/2309-4923-2022-2-4-9Старовойтов В.В., Ахунджанов У.Ю. Новый признак для описания изображений рукописной подписи на базе локальных бинарных шаблонов / В.В. Старовойтов, У.Ю. Ахунджанов // Информатика. – 2022. – Т. 19, № 3. – С. 6273. DOI: 10.37661/1816-0301-2022-19-3-62-73Starovoitov V.V., Akhundjanov U.Yu. A new feature for handwritten signature image description based on local binary patterns. Informatika [Informatics], 2022, vol. 19, no. 3, pp. 62-73 (In Russ.). DOI: 10.37661/1816-0301-2022-19-3-62-73

The authors declare that there are no conflicts of interest present.