Scientific and Technical Journal


ISSN Print 2221-3937
ISSN Online 2221-3805

The goal of this paper is to develop a method for the classification of region of document image as graphic or text content type. As an input of the method segment (rectangular region of an image) containing a homogeneous content – text or graphic object – is used. Region analysis is performed on the assumption that it is a text region, ​​projection partition allows to split this region into objects of smaller size. Construction of the narrative function characterizing areas of the image sub-elements is proposed. Feature size distribution for the next processing step is used for training and classification using support vector machines. Using the proposed approach allows to classify text area with a probability of 0.9859, graphics region – with a probability of about 0.9451. An investigation of the drawbacks and limitations of the proposed method was performed, areas of its applications were discovered. Scientific results of the paper can be used in automatic document image processing, analysis and pattern recognition.


1. Anupama N., Rupa Ch., and Sreenivasa E. Reddy. Character Segmentation for Telugu Image Document using Multiple Histogram Projections,(2013),Global Journal of Computer Science and Technology Graphics & Vision, Vol. 13, Issue 5, Version 1.0(In English).

2. Bloomberg D.S., Multiresolution Morpho-logical Approach to Document Image Analysis, (1991),Proceedings of  International Conference in Document Analysis and Recognition, pp. 963 –971(In English).

3. Bukhari, S.S., Azawi, M.I.A.A., Shafait, F., and Breuel T.M. Document Image Segmentation using Discriminative Learning over Connected Components, (2010),Proceeding of the International Workshop on Document Analysis Systems, ACM New York, NY, USA, pp. 183 – 190(In English).

4. Bukhari S.S., Shafait F., and Breuel T.M. Improved Document Image Segmentation Algorithm using Multiresolution Morphology, (2011), Proceedings of theXVIII Document Recognition and Retrieval Conference, San Jose, CA, USA, January 24– 29, 2011, pp.1 – 10(In English).

5. Gao J., Yang J., Zhang Y., and Waibel A., Text Detection and Translation from Natural Scenes, tech. report CMU-CS-01-139, (2001), Computer Science Department, Carnegie Mellon University(In English).

6. Gupta N., and Bange V.K., Image Segmentation for Text Extraction,(2012), Proceedings of the 2nd International Conference on Electrical, Electronics and Civil Engineering, Singapore, April  28– 29, 2012(In English).

7. Lazzara G., and Geraud T., Efficient Multiscale Sauvola's Binarization,(2013), International Journal of Document Analysis and Recognition(In English).

8. Le D.X., Thoma G.R., and Wechsler H. Automated Borders Detection and Adaptive Segmentation for Binary Document Images,(1996), Proceedings of the International Conference on Pattern Recognition,Vol. 7276, pp. 737 – 741(In English).

9. Lewis D., Agam G., Argamon S., Frieder O., Grossman D., and Heard J. Building a Test Collection for Complex Document Information Processing, (2006), Proceedings of the 29th Annual International ACM SIGIR Conference, pp. 665–666(In English).

10. Likforman-Sulem L., Zahour A., and Taconet B. Text Line Segmentation of Historical Documents: a Survey,(2007),International Journal on Document Analysis and Recognition, Springer, Vol. 9, Issue 2, pp.123 – 138(In English).

11. SVM Parameters, (2014), available at: (accessed 29 January 2014)(In English).

12. Szeliski R.Computer Vision: Algorithms and Applications, Springer-Verlag London Limited, 2011(In English).

Last download:
15 Jan 2020

[ © KarelWintersky ] [ All articles ] [ All authors ]
[ © Odessa National Polytechnic University, 2014-2018. Any use of information from the site is possible only under the condition that the source link! ]