350 rub
Journal Radioengineering №12 for 2015 г.
Article in number:
Detection and localization of text shapes in images
Keywords: image text shapes a text shapes localization second discrete difference support function
Authors:
A.V. Korennoi - Dr. Sc. (Eng.), Professor, Department of Communication facilities, MESC «Zhukovsky-Gagarin Air Force Academy» (Voronezh). E-mail: korennoj@mail.ru D.S. Yudakov - Lecturer, Department of Communication facilities, MESC «Zhukovsky-Gagarin Air Force Academy» (Voronezh). E-mail: smersh-2@mail.ru
Abstract:
Segmentation of the staff and location of the useful information from its content is the necessary requirement to functioning the modern information retrieval automatic system. Most of the automatic systems have the purpose of searching of the text information in the areas with the text information. Area of the image containing the text we will call as the text shapes. In this paper the new way of detection and localization of the text shapes in the images is presented. This method is based on develop-ment of a method of detection of horizontal differences of brightness. The procedure of detection of the text shapes is reduced to forma-tion of the sufficient statistics of the image and its comparison with a threshold. As a background we consider all the staff of the image taking into account that the text has a rather small area in comparison with the all image. As a result it does not affect the statistical characteristics of the whole image. We choose the value of a threshold using the Neyman-Pearson criteria (by a given probability of a false alarm) taking into account that the distribution of sufficient statistics corresponds to the so-called one-side normal distribution law. The proposed algorithm of detection and localization of the text shapes in the image has a simplicity in realization and allows to receive the characteristics not worse than the known algorithms.
Pages: 162-168
References

 

  1. Sysoeva D.A. Covremennye podkhody k poisku izobrazhenijj, soderzhashhikh tekst. Informacionnye upravljajushhie sistemy i kompjuternyjj monitoring (IUS KM-2013) // Materialy IV Mezhdunar. nauchno-tekhnich. konf. studentov, aspirantov i molodykh uchenykh. 2013. Doneck: DonNTU. T. 1. S. 387−391.
  2. Vasileva N. Obnaruzhenie teksta na izobrazhenijakh. nvassilieva@hp.com HP Labs Russia 8 aprelja 2012. Computer Science klub. Prezentacija:[EHlektronnyjj resurs]. URL: http://www. myshared.ru/slide/174588/. (Data obrashhenija: 20.05.2015).
  3. Aghajari G., Shanbehzadeh J., Sarrafzadeh A.A text localization algorithm in color image via new projection profile// Proceedings of the International MultiConference of Engineers and Computer Scientists 2010 Vol II.IMECS 2010. March 17−19. 2010. Hong Kong.
  4. Korennojj A.V. Matematicheskie modeli polutonovykh izobrazhenijj // Radiotekhnika. 2007. № 8.
  5. Prehtt U. Cifrovaja obrabotka izobrazhenijj / Per. s angl. M.: Mir. 1982. Kn.1. 312 s.
  6. Bogoslovskijj A.V. Obrabotka mnogomernykh signalov. M.: Radiotekhnika. 2013. 168 s.
  7. SHarov S.A. Statistika slov v russkom jazyke. [EHlektronnyjj resurs]. URL: http://www.lingvisto.org/artikoloj/ru_stat.html.(Data obrashhenija: 10.06.2015).
  8. Karpov I.G., Karpov M.G., Proskurin D.K. Metody obobshhennogo verojatnostnogo opisanija i identifikacii negaussovskikh sluchajjnykh velichin i processov. Voronezh: Izd-vo VGU. 2010. 172 s.
  9. Gonsales R., Vuds R. Cifrovaja obrabotka izobrazhenijj. M.:Tekhnosfera. 2005. 1072 s.
  10. JAne B. Cifrovaja obrabotka izobrazhenijj. M.:Tekhnosfera. 2007. 583 s.