In automatic numeral digit recognition system, feature collection is most important aspect for achieving high recognition performance. To attain this, we proposes model for printed numeral digit recognition using number of contours, skeleton features such as number of end points, number of horizental and vertical crossings Number of watersheds, and ratio between the number of foreground pixels in upper half-part and lower half-part of the numerical digit image. Based on these features the present study designed user defined classification algorithm for printed numerical digit recognition. To find the effectiveness of the proposed algorithm, these features are given as an input for standard classification algorithms like k–nearest neighbor classifier and other classification algorithms to evaluate the results. The experimental results prove that the proposed features are well suited for printed digit recognition for both user and standard classification algorithms. The novelty of the proposed method is size and shape invariant.

Structural ,Skeleton Features.K-nn,Classification,Watersheds,contours

[1] G. E. Hinton, S. Osindero, and Y.-W. Teh, “A fast learning algorithm for deep belief nets,” Neural Comput., vol. 18, no. 7, pp. 1527–1554, Jul. 2006.
[2] H. Lee, R. Grosse, R. Ranganath, and A. Y. Ng, “Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations,” in Proceedings of the 26th Annual International Conference on Machine Learning, ser. ICML ’09. New York, NY, USA: ACM, 2009, pp. 609–616.
[3] B. Wicht and J. Hennebert, “Camera-based sudoku recognition with deep belief network,” in Soft Computing and Pattern Recognition (SoCPaR), 2014 6th International Conference of, Aug 2014, pp. 83– 88.
[4] Chen, G., Li, Y. and Srihari, S.N. (2016) Word Recognition with Deep ConditionalRandom Fields. Proceedings of the IEEE International Conference on Image Processing(ICIP ), Phoenix, AZ, 25-28 September 2016, 1928-1932.
[5] arulla, G.A., Murru, N. and Rossini, R. (2016) A Fuzzy Approach for Segmentationof Touching Characters. arXiv:1612.04862v1. https://arxiv.org/abs/1612.04862
[6] Jaderberg, M., Simonyan, K., Vedaldi, A. and Zisserman, A. (2014) Reading Text inthe Wild with Convolutional Neural Networks. arXiv preprint arXiv:1412.1842
[7] Liu, Z., Li, Y., Ren, F. and Yu, H. (2016) A Binary Convolutional Encoder-DecoderNetwork for Real-Time Natural Scene Text Processing. arXiv:1612.03630v1.https://arxiv.org/abs/1612.03630v1
[8] Yin, X.C., Yin, X., Huang, K. and Hao, H.W. (2014) Robust Text Detection in NaturalScene Images. IEEE Transactions on Pattern Analysis and Machine Intelligence, 36, 970-983. https://doi.org/10.1109/TPAMI.2013.182
[9] Oivind Trier, Anil Jain, Torfiinn Taxt, ―A feature extraction method for character recognition-A survey pattern Recg, vol 29, No 4, pp641-662, 1996
[10] P.Nagabhushan, S.A.Angadi, B.S.Anami, ―A fuzzy statistical approach of Kannada Vowel Recognition based on Invariant Moments‖, Proc. Of NCDAR-2003, Mandy, Karnataka, India, pp275-285, 2003
[11] G. G. Rajput, Rajeswari Horakeri, Sidramappa Chandrakant “Printed and Handwritten Kannada NumeralRecognition Using Crack Codes and Fourier Descriptors Plate” RTIPPR, 2010. IJCA Special Issue on “RecentTrends in Image Processing and Pattern Recognition” pp.53-58
[12] Graham, D.N “Image transmission by two-dimensional contour coding” Proceedings of the IEEE Volume: 55,Issue: 3 Pages: March 1967 336 – 346
[13] U. Pal, T. Wakabayashi, N. Sharma and F. Kimura,"Handwritten Numeral Recognition of Six Popular IndianScripts", In Proc. 9th International Conference on Document Analysis and Recognition. pp. 749-753, Curitiba,Brazil, September 24-26, 2007
[14] Seong- W han Lee, "Multi layer cluster neural network for totally unconstrained handwriritten numeralrecognition. Neural Networks. Vo I. 8. No 5. pp.409-4 1 8, 1984
[15] Image Processing Techniques For Machine Vision [online].Available:http://www.eng.fiu.edu/me/robotics/elib/am_st_fiu_ppr_2000.pdf