Improving Offline Handwritten Digit Recognition Using Concavity-Based Features

  • Miran Karic J. J. Strossmayer University of Osijek Croatia, 31000 Osijek, Kneza Trpimira 2b
  • Goran Martinovic


This paper examines benefits of using concavity-based structural features in recognition of handwritten digits. An overview of existing concavity features is presented and a new method is introduced. These features are used as complementary features to gradient and chaincode features, both among the best performing features in handwritten digit recognition. Two support vector classifiers (SVCs) are chosen for classification task as the top performers in previous works; SVC with radial basis function (RBF) kernel and the SVC with polynomial kernel. For reference, we also used the k-nearest neighbor (k-NN) classifier. Results are obtained on MNIST, USPS and DIGITS datasets. We also tested dataset independency of various feature vectors by combining different datasets. The introduced feature extraction method gives the best results in majority of tests.


[1] C.-L. Liu, K. Nakashima, H. Sako, H. Fujisawa, Handwritten digit recognition: investigation of normalization and feature extraction techniques, Pattern Recognition, 37(2):265-279, 2004.

[2] C.-L. Liu, K. Nakashima, H. Sako, H. Fujisawa, Handwritten digit recognition: benchmarking of state-of-the-art techniques, Pattern Recognition, 36(10):2271-2285, 2003.

[3] M. H. Nguyen, F. de la Torre, Optimal feature selection for support vector machines, Pattern Recognition, 43(3):584-591, 2010.

[4] U. Kressel, J. SchĂźrmann, Pattern classification techniques based on function approximation, Handbook of Character Recognition and Document Image Analysis, 49-78, 1997.

[5] B. P. Chacko, P. Babu Anto, Comparison of Statistical and Structural Features for Handwritten Numeral Recognition, Proc. of the Int. Conf. on Computational Intelligence and Multimedia Applications (ICCIMA 2007), Washington, DC (USA), 296-300, 2007.

[6] H. Liu, X. Ding, Handwritten Character Recognition Using Gradient Feature and Quadratic Classifier with Multiple Discrimination Schemes, Proc. of the Eighth Int. Conf. on Document Analysis and Recognition (ICDAR '05), Washington, DC (USA), 19-25, 2005.

[7] O. D. Trier, A. K. Jain, T. Taxt, Feature Extraction Methods for Character Recognition - A Survey, Pattern Recognition, 29(4):641-662, 1996.

[8] G. Vamvakas, B. Gatos, I. Pratikakis, N. Stamatopoulos, A. Roniotis, S. J. Perantonis, Hybrid off-line OCR for isolated handwritten Greek characters, Proc. of the Fourth IASTED Int. Conf. on Signal Processing, Pattern Recognition, and Applications, Innsbruck (Austria), 197-202, 2007.

[9] J. Favata, G. Srikantan, S. Srihari, Handprinted character/digit recognition using a multiple feature/resolution philosophy, Fourth International Workshop on Frontiers in Handwriting Recognition, Taipei (Taiwan), 67-70, 1994.

[10] Y. LeCun, L. Bottou, Y. Bengio, P. Haffner, Gradient-Based Learning Applied to Document Recognition, Proceedings of the IEEE, 86(11):2278-2324, 1998.

[11] J. J. Hull, A Database for Handwritten Text Recognition Research, Pattern Analysis and Machine Intelligence, 16(5):550-554, 1993.

[12] A. K. Seewald, Digits - A Dataset for Handwritten Digit Recognition, Austrian Research Institut for Artificial Intelligence Technical Report, Vienna (Austria), 2005.

[13] C. Cortes, V. Vapnik, Support-Vector Networks, Machine Learning, 20(3):273-297, 1995.

[14] L. Van der Maaten, A New Benchmark Dataset for Handwritten Character Recognition, Tilburg University Technical Report, 2009.

[15] A. K. Seewald, On the Brittleness of Handwritten Digit Recognition Models, Technical Report, Seewald Solutions, Vienna (Austria), 2009.

[16] M. Karic, Concavity paper source code. [Online] Cited 2011-08-30. Available at:

[17] R. V. D. Heiden, F. C. A. Gren, The Box-Cox metric for nearest neighbor classification improvement, Pattern Recognition, 30(2):273-279, 1997.

[18] C. B. Barber, D. P. Dobkin, H. T. Huhdanpaa, The Quickhull Algorithm for Convex Hulls, ACM Trans. on Mathematical Software, 22(4):469-483, 1996.

[19] C.-C. Chang, C.-J. Lin, LIBSVM: a library for support vector machines, ACM Trans. on Intelligent Systems and Technology, 2(3): 1-39, 2012. Software available at

[20] N. Otsu, A Threshold Selection Method from Gray-Level Histograms. IEEE Trans. on Systems, Man and Cybernetics, 9(1):62-66, 1979.

[21] M. R. Gupta, N. P. Jacobson, E. K. Garcia, OCR binarization and image pre-processing for searching historical documents, Pattern Recognition, 40(2):389-397, 2007.

[22] Y. LeCun, The MNIST database of handwritten digits. [Online] Cited 2011-08-30. Available at:

[23] D. Keysers, Experimental results on the USPS database. [Online] Cited 2011-08-30. Available at:

[24] C. E. Rasmussen, C. K. I. Williams, Gaussian Processes for Machine Learning, 2nd ed., The MIT Press, 2006.
How to Cite
KARIC, Miran; MARTINOVIC, Goran. Improving Offline Handwritten Digit Recognition Using Concavity-Based Features. INTERNATIONAL JOURNAL OF COMPUTERS COMMUNICATIONS & CONTROL, [S.l.], v. 8, n. 2, p. 220-234, feb. 2013. ISSN 1841-9844. Available at: <>. Date accessed: 12 july 2020. doi:


Complementary features, concavity features, digit recognition, feature extraction, handwritten character recognition, off-line recognition.