A projection-based criterion for polish typewritten/handwritten word identification using a new approach in segmentation

• Autorzy: Saeed K. , Predko A. , Rybnik M.
• Języki publikacji: EN

Abstrakty

EN

This paper presents experiments on Polish scripts recognition by a method based on projection method fused with Toeplitz matrix approach. A new approach in segmentation of letters is presented. The essential part of this work is based on modification of known segmentation methods. The observations have allowed us to create a new criterion for word segmentation. Projection approach (which has been earlier used to identify Arabic and Latin signs and gives good results) is used to identify Polish letters. Toeplitz approach is also applied to allow preparing data for further processing and analysis. Samples of typewritten and handwritten words were segmented and identified in order to validate the approach. The obtained results are satisfying and encourage further periments.

Słowa kluczowe

EN

segmentation; overlapping problem; Toeplitz approach

PL

segmentacja

• Wydawca: Wydawnictwo Politechniki Częstochowskiej
• Czasopismo: Computing, Multimedia and Intelligent Techniques
• Rocznik: 2006
• Tom: Vol. 2, nr 1
• Strony: 25--34
• Opis fizyczny: Bibliogr. 13 poz., rys., tab.

Twórcy

autor

Saeed K.

autor

Predko A.

autor

Rybnik M.

• Białystok Technical University, Faculty of Computer Science Computer Engineering Department, ul. Wiejska 45A, 15-351 Białystok, Poland,

Bibliografia

• [1] K. Saeed, Image Analysis for Object Recognition, Bialystok Technical University Press, Bialystok 2004.
• [2] T.S. El-Sheikh, R. Guindi, Computer recognition of Arabic cursive scripts, Pattern Recognition, 21,4, 293-302, 1988.
• [3] H. Almuallim, S. Yamaguchi, A method of recognition of Arabic cursive handwriting, IEEE Transactions on Pattern Analysis and Machine Intelligence, 14, 8, 853-857,1987.
• [4] F. El-Khaly, M. Sid-Ahmed, Machine recognition of optically captured machine printed Arabic text, Pattern Recognition, 23, 11, 1207-1214, 1990.
• [5] K, Saeed. M. Rybnik, M. Tabędzki, Implementation and Advanced Results on the Non-Interrupted Skeletonization Algorithm, LNCS, ed. W. Skarbek, 2124, 601-609, Springer-Verlag, Heidelberg 2001.
• [6] K. Saeed, M. Tabędzki, M. Adamski, A New Approach for Object-Feature Extract and Recognition, Proc. 9th Intern. Conf. on Advanced Computer Systems - ACS'02, Miedzyzdroje, 23-25 October 2002, 389-397.
• [7] D. Burr, Experiments on Neural Net Recognition of Spoken and Written Text, IEEE Trans. On Acoustic, Speech and Signal Proc., 36, 7, 1162-1168, 1988.
• [8] K. Saeed, A Projection Approach for Arabic Handwritten Characters Recognition, Quo Vadis Computational Intelligence? New Trends and Apps. in Computational Intelligence, eds. P. Sincak, J. Vascak, 106-111, Physica-Verlag, Berlin 2000.
• [9] P. Jaskulski, Digital Taxonomy. Introduction to Classification Problematics, Biuletyn Antropologiczny, 1, Warszawa 1997 (in Polish).
• [10] M. Kurzynski, Object recognition, Oficyna Wydawnicza Politechniki Wroc³awskiej, Wroc³aw 1997 (in Polish).
• [11] R.A. Johnson, Dean W. Wichern, Applied Multivariate Statistical Analysis, Prentice-Hall, 1991.
• [12] M. Rybnik, Contribution to the Modelling and the Exploitation of Hybrid Multiple Neutral Networks Systems: Application to Intelligent Processing of Information (doctoral thesis), University Paris XII - Val de Marne, Creteil 2005.
• [13] A. Amin, J. Mari: Machine Recognition and Correction of Printed Arabic Text, IEEE Trans, on Systems Man, and Cybernetics, 19, 5, 1300-1306, 1989.