Abstract. 

This paper proposes a modified version of the Viola-Jones algorithm to improve the accuracy of object localization in document images when searching for and locating a stamp. The modification involves expanding the feature space with square centrally symmetric features, resulting in a significant increase in accuracy. The effectiveness of this approach is demonstrated through its application to locating the official stamp on Russian Federation passports, and the paper provides quantitative estimates of the improvement achieved.

Keywords: 

Viola-Jones cascades, machine learning, image processing, document recognition, forensics, pattern detection.

PP. 31-39.

DOI: 10.14357/20790279230304

 

References

1. Viola P., Jones M.J. Robust Real-Time Face Detection. International Journal of Computer Vision 57, P. 137–154 (2004). https://doi.org/10.1023/ B:VISI.0000013087.49260.fb

2. Huang J., Shang Y., Chen H. Improved Viola-Jones face detection algorithm based on HoloLens. J Image Video Proc. 2019. V. 2019. no. 1. P. 1-11.

3. Lu Wen-Yao and Y.A.N.G. Ming. Face detection based on viola-jones algorithm applying composite features. 2019 International Conference on Robots & Intelligent System (ICRIS) P. 82-85.

4. Polyakov I.V., Kuznetsova E.G., Usilin S.A., Nikolaev D.P. Postroenie optimalnykh kaskadov violy–dzhonsa pri pomoshchi “zhadnykh” algoritmov perebora upravlyayushchikh parametrov s promezhutochnym kontrolem po validatsionnoi vyborke [Training optimal viola–jones detectors using greedy algorithms for selecting control parameters with intermediate validation on each level]. Sensornye sistemy [Sensory systems]. 2016. V. 30(3). P. 241-248 (in Russian).

5. Pchelincev S.Y., Arzamascev A.A. Primenenie Metoda Violy-Dzhonsa dlya raspoznavaniya dorozhnyh znakov [Applying the Viola-Jones Method to Traffic Sign Recognition]. Rossijskaya nauka v sovremennom mire [Russian science in the modern world]. 2019. P. 111-112.

6. Jose A, Thodupunoori H, Nair BB. A novel traffic sign recognition system combining Viola–Jones framework and deep learning. InSoft Computing and Signal Processing. Springer, Singapore. 2019. P. 507-517.

7. Hussein I.J. et al. Fully automatic segmentation of gynaecological abnormality using a new viola– jones model. Comput. Mater. Contin. 2021. V. 66. P. 3161-3182.

8. Usilin S.A., Slavin O.A., Arlazarov V.V. Memory Consumption and Computation Efficiency Improvements of Viola–Jones Object Detection Method for Remote Sensing Applications. Pattern Recognition and Image Analysis. 2021. V. 31. No. 3. P. 571-579.

9. Xu Y, Yu G, Wang Y, Wu X, Ma Y. A hybrid vehicle detection method based on viola-jones and HOG+ SVM from UAV images. Sensors. 2016. Aug 19;16(8):1325.

10. Usilin S. et al. Visual appearance based document image classification. Proceedings - International Conference on Image Processing, ICIP. Hong Kong. 2010. P. 2133–2136.

11. Arlazarov V.V., Nikolaev D.P., Usilin S.A. and Sholomov D.L. Raspoznavanie gil’oshirnyh elementov: opredelenie stranic pasporta RF [Guilloche Elements Recognition Applied to Passport Page Processing]. Trudy ISA RAN. 2013. V. 63. No 3. P. 106-110.

12. Arlazarov V.V., Matalov D.P. and Usilin S.A. Lokalizaciya obraza pechati na dokumente, udostoveryayushchem lichnost’, metodom mashinnogo obucheniya [Localization of the seal on the identity document image using machine learning approach]. Trudy ISA RAN. Special Issue. 2018. No. 1. Vol. 68. P. 158-166.

13. Kuznetsova E., Shvets E., Nikolaev D. Viola- Jones based hybrid framework for real-time object detection in multispectral images. ICMV 2015 V. 9875. 2015. P. 98750N1-98750N6. DOI: 10.1117/12.2228707.

14. Aliev M.A., Arlazarov V.V., Matalov D.P., Nikolaev D.P., Polevoy D.V. and Usilin S.A. Sposob ocenki dejstvitel’nosti dokumenta pri pomoshchi opticheskogo raspoznavaniya tekstana izobrazhenii kruglogo ottiska pechati/shtampa na cifrovom izobrazhenii dokumenta [Method of assessing the validity of a document using optical recognition of text on the image of a round impression of a print / stamp on a digital image of a document] // Rospatent RU 2750395 C1.

15. Everingham M., Van Gool L., Williams C.K.I. et al. The PASCAL Visual Object Classes (VOC) Challenge. Int J Comput Vis. 2010. V. 88. No. 2, P. 303–338. https://doi.org/10.1007/s11263-009- 0275-4