Performance Evaluation of a Recognition System on the VLIW Architecture by the Example of the Elbrus Platform

Abstract

This paper overviews modern computing devices based on the Elbrus VLIW architecture and presents experimental results for the performance evaluation of the Smart IDReader document recognition system on these devices. Methods for speeding up the recognition system on the Elbrus platform are described, experimental estimates of the speedup improvements are presented, and the performance of various devices based on Elbrus processors is compared.

This is a preview of subscription content, log in to check access.

REFERENCES

  1. 1

    Mastov, A., Konovalenko, I., and Grigoryev, A., Application of random ferns for non-planar object detection, Proc. 8th Int. Conf. Machine Vision, 2015.

  2. 2

    Kopenkov, V.N. and Myasnikov, V.V., Detection and tracking of vehicles based on the video-registration information, Proc. 23rd Int. Conf. Computer Graphics, Visualization, and Computer Vision (WSCG), Plzen, 2015, pp. 65–69.

  3. 3

    Arlazarov, V.V., Cognitive Form: A distributed system for stream recognition of standard forms of documents, Tr. konf. “Raspoznavanie obrazov i analiz izobrazhenii: novye informatsionnye tekhnologii” (ROAI) (Proc. Conf. Pattern Recognition and Image Analysis: New Information Technologies), Novgorod, 2002, pp. 41–45.

  4. 4

    Arlazarov, V., Postnikov, V., and Sholomov, D., Cognitive Forms: A system for mass recognition of structured documents, Sb. tr. Inst. Sistemnogo Anal. Ross. Akad. Nauk (Proc. Inst. Syst. Anal. Russ. Acad. Sci.), Moscow: URSS, 2002, pp. 35–46.

    Google Scholar 

  5. 5

    Omondi, A.R. and Rajapakse, J.C., FPGA Implementations of Neural Networks, New York: Springer, 2006.

    Google Scholar 

  6. 6

    Farabet, C., Poulet, C., Han, J.Y., and LeCun, Y., CNP: An FPGA-based processor for convolutional networks, Proc. Int. Conf. Field Programmable Logic and Applications, Prague, 2009, pp. 32–37.

  7. 7

    Yang, Q., Wang, C., Gao, Y., Qu, H., and Chang, E.Y., Inertial sensors aided image alignment and stitching for panorama on mobile phones, Proc. 1st Int. Workshop Mobile Location-Based Service (MLBS), New York, 2011, pp. 21–30.

  8. 8

    Luqman, M.M., Gomez-Kramer, P., and Ogier, J.-M., Mobile phone camera-based video scanning of paper documents, Revised Selected Papers (Proc. 5th Int. Workshop Camera-Based Document Analysis and Recognition (CBDAR), Washington, 2013), Iwamura, M. and Shafait, F., Eds., 2014, pp. 164–178.

  9. 9

    Ramisa, A., Tapus, A., De Mantaras, R.L., and Toledo, R., Mobile robot localization using panoramic vision and combinations of feature region detectors, Proc. IEEE Int. Conf. Robotics and Automation (ICRA), Pasadena, 2008, pp. 538–543.

  10. 10

    Kim, A.K., Bychkov, I.N., et al., Russian technologies “Elbrus” for personal computers, servers, and supercomputers, Sovrem. Inf. Tekhnol. IT Obraz., 2014, no. 10, pp. 39–50.

  11. 11

    Kim, A.K., Perekatov, V.I., and Ermakov, S.G., Mikroprotsessory i vychislitel’nye kompleksy semeistva “El’brus” (Microprocessors and Computing Complexes of the Elbrus Family), St. Petersburg: Piter, 2013.

  12. 12

    Ishin, P.A., Loginov, V.E., and Vasil’ev, P.P., Computation speedup using high-performance mathematical and multimedia libraries for the Elbrus architecture, Vestn. Vozdushn.-Kosm. Oborony, 2015, vol. 8, no. 4, pp. 64–68.

    Google Scholar 

  13. 13

    Bezmaternykh, P., Nikolaev, D., and Arlazarov, V.L., Textual blocks rectification method based on fast Hough transform analysis in identity documents recognition, Proc. 10th Int. Conf. Machine Vision (ICMV), Vienna, 2017.

  14. 14

    Stahlberg, F. and Vogel, S., The QCRI recognition system for handwritten Arabic, Proc. Int. Conf. Image Analysis and Processing, 2015, pp. 276–286.

  15. 15

    Chernov, T.S., Razumnyi, N.P., Kozharinov, A.S., and Arlazarov, V.V., Quality estimation of input images in video stream recognition systems, Inf. Tekhnol. Vychisl. Sist., 2017, no. 4, pp. 71–82.

  16. 16

    Usilin, S., Nikolaev, D., Postnikov, V., and Schaefer, G., Visual appearance based document image classification, Proc. Int. Conf. Image Processing (ICIP), 2010, pp. 2133–2136.

  17. 17

    Bocharov, N.A., Limonova, E.E., Paramonov, B.N., and Usilin, S., Optimization of a modified Viola–Jones method for the Elbrus computing architecture, Tr. Inst. Sistemnogo Anal. Ross. Akad. Nauk (Proc. Inst. Syst. Anal. Russ. Acad. Sci.), 2017, vol. 67, no. 4, pp. 10–21.

  18. 18

    Chen, N. and Blostein, D., A survey of document image classification: Problem statement, classifier architecture, and performance evaluation, Int. J. Doc. Anal. Recognit., 2007, vol. 10, no. 1, pp. 1–16.

    Article  Google Scholar 

  19. 19

    Vorontsov, K.V., Additive regularization for topic models of text collections, Dokl. Math., 2014, vol. 89, no. 3, pp. 301–304.

    MathSciNet  Article  MATH  Google Scholar 

  20. 20

    Junga, K., Kimb, K.I., and Jain, A.K., Text information extraction in images and video: A survey, Pattern Recognit., 2004, pp. 977–997.

  21. 21

    Slugin, D.G. and Arlazarov, V.V., Detection of text fields in documents by using image processing methods, Tr. Inst. Sistemnogo Anal. Ross. Akad. Nauk (Proc. Inst. Syst. Anal. Russ. Acad. Sci.), 2017, vol. 67, no. 4, pp. 65–73.

  22. 22

    Limonova, E., Ilin, D., and Nikolaev, D.P., Improving neural network performance on SIMD architectures, Proc. 8th Int. Conf. Machine Vision (ICMV), 2015, pp. 1–6.

  23. 23

    Limonova, E., Sheshkus, A., and Nikolaev, D., Computational optimization of convolutional neural networks using separated filters architecture, Int. J. Appl. Eng. Res., 2016, vol. 11, no. 11, pp. 7491–7494.

    Google Scholar 

  24. 24

    Limonova, E., Sheshkus, A., Nikolaev, D.P., Ivanova, A., Il’in, D., and Arlazarov, V.L., Efficiency optimization of the first layers of deep convolutional neural networks, Vestn. RFFI, 2016, no. 4, pp. 84–96.

  25. 25

    Limonova, E., Sheshkus, A., Ivanova, A., and Nikolaev, D.P., Convolutional neural network structure transformations for complexity reduction and speed improvement, Pattern Recognit. Image Anal., 2018, no. 1, pp. 24–33.

  26. 26

    Ilin, D., Limonova, E., Arlazarov, V.V., and Nikolaev, D., Fast integer approximations in convolutional neural networks using layer-by-layer training, Proc. 9th Int. Conf. Machine Vision (ICMV), 2017, pp. 1–5.

  27. 27

    Gavrikov, M.B., Pestryakova, N.V., Slavin, O.A., and Farsobina, V.V., Development of the polynomial regression method and its practical application to recognition problems, Preprint of Keldysh Inst. Appl. Math., 2006, no. 25, pp. 1–22.

  28. 28

    Krizhevsky, A., Sutskever, I., and Hinton, G.E., Imagenet classification with deep convolutional neural networks, Advances in Neural Information Processing Systems, 2012.

  29. 29

    Manzhikov, T.V., Slavin, O.A., Faradjev, I.A., and Janiszewski, I.M., Algorithms for correcting recognition results using N-grams, Pattern Recognit. Image Anal., 2007, vol. 4, no. 27, pp. 1054–6618. https://doi.org/10.1134/S1054661817040125

    Google Scholar 

  30. 30

    Bulatov, K., Manzhikov, T., Slavin, O., Faradjev, I., and Janiszewski, I., Trigram-based algorithms for OCR result correction, 2017, vol. 10341, p. 1034100. https://doi.org/10.1117/12.2268559

  31. 31

    Llobet, R., Navarro-Cerdan, J.R., Perez-Cortez, J.-C., and Arlandis, J., OCR post-processing using weighted finite state transducers, Proc. 12th Int. Conf. Frontiers in Hadwriting Recognition, Kolkata, 2010, pp. 2021–2024.

  32. 32

    Bulatov, K.B., Selecting the optimal strategy for combining the results of frame-by-frame symbol recognition in video stream, Inf. Tekhnol. Vychisl. Sist., 2017, no. 3, pp. 45–55.

  33. 33

    Fiscus, J.G., A post-processing system to yield reduced word error rates: Recognizer output voting error reduction (ROVER), Proc. IEEE Workshop Automatic Speech Recognition and Understanding, 1997, pp. 347–354.

  34. 34

    Bulatov, K., Arlazarov, V., Chernov, T., Slavin, O., and Nikolaev, D., Smart IDReader: Document recognition in video stream, Proc. 14th IAPR Int. Conf. Document Analysis and Recognition (ICDAR), Kyoto, Japan, 2017, pp. 39–44.

  35. 35

    Arlazarov, V.V., Arlazarov, V.L., Bulatov, K., Nikola-ev, D.P., Polevoi, D., and Slavin, O., System for document recognition in video stream, Russian Federation’s State Register of Useful Models.

  36. 36

    Programming system for Elbrus platforms and MTsST-R. http://www.mcst.ru/sistema_programmirovaniya.

  37. 37

    Intel Threading Building Blocks (Intel TBB). http:// www.threadingbuildingblocks.org.

Download references

ACKNOWLEDGMENTS

This work was supported by the Russian Foundation for Basic Research, project nos. 17-29-03170, 17-29-03297, and 18-07-01384.

Author information

Affiliations

Authors

Corresponding authors

Correspondence to E. E. Limonova or N. A. Bocharov or N. B. Paramonov or D. S. Bogdanov or V. V. Arlazarov or O. A. Slavin or D. P. Nikolaev.

Additional information

Translated by Yu. Kornienko

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Limonova, E.E., Bocharov, N.A., Paramonov, N.B. et al. Performance Evaluation of a Recognition System on the VLIW Architecture by the Example of the Elbrus Platform. Program Comput Soft 45, 12–17 (2019). https://doi.org/10.1134/S0361768819010055

Download citation