Advertisement

Multimedia Tools and Applications

, Volume 78, Issue 21, pp 30027–30037 | Cite as

Study on the influence of variable stride scale change on image recognition in CNN

  • Chen Guo
  • Yue-lan LiuEmail author
  • Xuan Jiao
Article
  • 90 Downloads

Abstract

After the research based on the progressing image classification recognition method of CNN, the paper aims at the problem that the size of feature size of output map of image with different complexity cannot be well solved by the constant value stride. We bring up the idea which based on the variable stride length for constraint parameters to selectively select the size of the stride. It is helpful to improve the efficiency of selective extraction and recognition of important features. Later studies have proved that the deficiency issue of complex image characteristic extraction due to the large stride size could be averted by adopting the variable stride length method based on constraint parameters. In the meantime, the method also avoids low recognition efficiency due to the image complexity is sparse and, also, the stride size of the image is too small. The theoretically calculated results are in good agreement with the experimental results.

Keywords

CNN Stride Characteristic mapping Multimedia 

Notes

Acknowledgements

The Project was supported by the Natural Science Foundation of Liaoning Province (Grant No. 20170540131), Nature Science Foundation of Heilongjiang Province (Grant No.C201437), Natural Science Foundation of Heilongjiang Province (Grant No. QC2018082) and Basic Scientific Research Funds of Heilongjing Provincial Higher Education lnstitutions (Grant No. 2017-KYYWF-0140). And we wish to thank the anonymous reviewers who helped to improve the quality of the paper.

References

  1. 1.
    Bengio Y (2009) Learning deep architectures for AI[J]. foundations and trends in. Mach Learn 2(1):1–127MathSciNetCrossRefGoogle Scholar
  2. 2.
    Boureau YL, Ponce J, Lecun Y (2010) A theor-etical analysis of feature pooling in visual recognition [C]. Proc of International Conference on Machine Learning, p 111–118,Google Scholar
  3. 3.
    Chan T-H, Jia K, Gao S et al (2015) PCANet:a simple deep learning baseline for image classification[J]. IEEE Trans Image Process 24(12):5017–5032MathSciNetCrossRefGoogle Scholar
  4. 4.
    Dalal N, Triggs B (2005) Histograms of oriented gradients for human detection[C]. Computer Vision and Pattern Recognition, 2005. IEEE Computer Society Conference on Piscataway, NJ: IEEE, p 886–893Google Scholar
  5. 5.
    Fei-Fei L, Karpathy A, Johnson J (2016) CS231n:Convolutional neural networks for visual recognition. StanfordGoogle Scholar
  6. 6.
    Goodfellow I, Bengio Y, Courville A (2016) Deeplearning. Book in preparation for MIT PressGoogle Scholar
  7. 7.
    Han FY (2010) Analysis of modern multimedia technology features and key technologies. JCNU(NS) 29(3):129–131Google Scholar
  8. 8.
    Hinton GE, Salakhutdinov RR (2006) Reducing the dimensionality of data with neural networks. Science 313(5786):504–507MathSciNetCrossRefGoogle Scholar
  9. 9.
    Hinton GE, Osindero S, Teh YW (2006) A fast learning algorithm for deep belief nets [J]. Neural Comput 18(7):1527–1554MathSciNetCrossRefGoogle Scholar
  10. 10.
    Kalchbrenner N, Grefenstette E, Blunsom P (2014) A con-volutional neural network for modeling sentences. arXiv preprint arXiv:1404.2188Google Scholar
  11. 11.
    Koutník J, Greff K, Gomez F, et al (2014) A Cl-ockwork RNN [J]. Computer Science:1863–1871Google Scholar
  12. 12.
    Krizhevsky A, Hinton G (2009) Learning multiple layers of features from tiny imagesGoogle Scholar
  13. 13.
    Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks [C] . Proc of the Conf of Advances in Neural Information Processing Systems. Rostrevor, Ireland: Curran Associates, Inc, p 1097–1105Google Scholar
  14. 14.
    Le QV, Ranzato M, Monga R, et al (2013) Building high-level features using large scale unsupervised learning [C]. Proc of the IEEE Int Conf on Acoustics, Speech and Signal Processing. Piscataway, NJ: IEEE, p 8595–8598Google Scholar
  15. 15.
    Lecun Y, Boser B, Denker JS et al (2014) Backpropagation applied to handwritten zip code recognition [J]. Neural Comput 1(4):541–551CrossRefGoogle Scholar
  16. 16.
    Lee H, Grosse R, Ranganath R, et al (2009) Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations [C]. Proc of the 25th Annual Int Conf on Machine Learning. Piscataway, NJ: IEEE, p 609–616Google Scholar
  17. 17.
    Dan Meng (2017) Research on image classification method based on deep learning. East China Normal University Press, p 5–15Google Scholar
  18. 18.
    Qian Y (2012) An alternative algorithm for finding the maximum value of a function--golden section search method. JEIJP 28(12):140–142Google Scholar
  19. 19.
    Russakovsky O, Deng J, Su H et al (2015) Imagenet large scale visual recognition challenge [J]. Int J Comput Vis 115(3):211–252MathSciNetCrossRefGoogle Scholar
  20. 20.
    Wang B (2015) Research on image classification and image retrieval based on visual features and machine learning [D]. XiDian University, p 1–4Google Scholar
  21. 21.
    Yang J, Yu K, Gong Y, et al (2009) Linear spatial pyramid matching using sparse coding for image classification [J]. 1794–1801Google Scholar
  22. 22.
    Ye XY, Qin J (2006) The characteristics of shallow learning and deep learning are shown in the table. Educational technology guide. No.1, p 19–21Google Scholar
  23. 23.
    Yu K (2013) Large-scale deep learning at Baidu [C]. Proc of ACM International Conferenceon Information & Knowledge Management. p 2211–2212Google Scholar
  24. 24.
    Zhang C, Li X, Yan J, et al (2014) Sufficient statistics feature mapping over deep Boltzmann machine for detection[C]. International Conference on Pattern Recoginition (ICPR), p 827–832Google Scholar
  25. 25.
    Zhang J, Wang H, Yang G, Xiao H (2018) Review of deep learning. ARC 35(7):1–2Google Scholar
  26. 26.
    Jun Zhu (2015) Image classfication based on deep learning models. Ningbo University. p 6–7Google Scholar

Copyright information

© Springer Science+Business Media, LLC, part of Springer Nature 2018

Authors and Affiliations

  1. 1.Software Technology InstituteDalian Jiaotong UniversityDalianChina
  2. 2.College of Computer ScienceHarbin Normal UniversityHarbinChina
  3. 3.School of Information and Business ManagementDalian Neusoft University of InformationDalianChina

Personalised recommendations