Advertisement

A modified faster region-based convolutional neural network approach for improved vehicle detection performance

  • Qinghui Zhang
  • Chenxia Wan
  • Weiliang Han
Article
  • 10 Downloads

Abstract

Presently available algorithms employed for vehicle detection exhibit three main disadvantages: slow detection speed, poor small objects detection, and low detection precision. To solve above problems, the present work proposes a vehicle detection approach employing a modified faster region-based convolutional neural network (R-CNN). Firstly, this approach introduces a deep CNN-based on VGG-16 and inception architecture, and adds a set of convolutional kernels with a 1 × 1 size, called a deep convolutional network (DCN). Then, an accurate vehicle region network (AVRN) and a vehicle attribute learning network (VALN) are designed. The AVRN accurately generates vehicle-like regions in real time, and the VALN detects the corresponding classifications and locations of vehicle-like regions. To improve the detection precision, we introduce corresponding loss functions for the AVRN and VALN. The calculation speed is increased by alternately optimizing and jointly training the AVRN and VALN. Experimental results demonstrate that the modified faster R-CNN approach improves significantly vehicle detection performance relative to existing algorithms, where, compared to the standard state-of-the-art faster R-CNN vehicle detection approach, the mean average precision of the test results obtained by the modified approach is increased by 11% and the detection time is reduced by one-third.

Keywords

Convolutional neural network Faster R-CNN Modified inception model Vehicle detection 

Notes

Acknowledgements

This work is supported by the National Natural Science Foundation of China (No. U1404617), Outstanding Youth Project of Science and Technology Innovation Talent Program of Henan Province (No.174100510011), and Program for Innovative Research Team (in Science and Technology) in Henan Province University (No.16IRTSTHN026).

References

  1. 1.
    Adankon MM, Cheriet M (2002) Support vector machine. Comput Sci 1 (4):1–28.  https://doi.org/10.1007/978-0-387-73003-5-299 CrossRefzbMATHGoogle Scholar
  2. 2.
    Bouchard G (2011) Clustering and classification employing softmax function including efficient boundsGoogle Scholar
  3. 3.
    Chen Z, Wang C, Wen C et al (2015) Vehicle detection in high-resolution aerial images via sparse representation and superpixels. IEEE Trans Geosci Remote Sens 54(1):103–116CrossRefGoogle Scholar
  4. 4.
    Cheng G (2016) A survey on object detection in optical remote sensing images. ISPRS J Photogramm Remote Sens 117:11–28.  https://doi.org/10.1016/j.isprsjprs.2016.03.014 CrossRefGoogle Scholar
  5. 5.
    Cheng G, Zhou P, Han J (2016) Learning rotation-invariant convolutional neural networks for object detection in vhr optical remote sensing images. IEEE Trans Geosci Remote Sens 54(12):7405–7415CrossRefGoogle Scholar
  6. 6.
    Cheng HY, Weng CC, Chen YY (2012) Vehicle detection in aerial surveillance using dynamic bayesian networks. IEEE Trans Image Process 21(4):2152–2159MathSciNetCrossRefGoogle Scholar
  7. 7.
    Dollr P, Appel R, Belongie S et al (2014) Fast feature pyramids for object detection. IEEE Trans Pattern Anal Mach Intell 36(8):1532–1545.  https://doi.org/10.1109/TPAMI.2014.2300479 CrossRefGoogle Scholar
  8. 8.
    Fan Q, Brown L, Smith J (2016) A closer look at faster r-cnn for vehicle detection. Intelligent Vehicles Symposium, pp 124–129Google Scholar
  9. 9.
    Girshick R (2015) Fast r-CNN. In: IEEE international conference on computer vision, pp 1440–1448, Santiago, ChileGoogle Scholar
  10. 10.
    Girshick R, Donahue J, Darrell T et al (2014) Rich feature hierarchies for accurate object detection and semantic segmentation. In: IEEE conference on computer vision and pattern recognition, pp 580–587. Columbus, OH, USAGoogle Scholar
  11. 11.
    Girshick R, Donahue J, Darrell T, et al. (2016) Region-based convolutional networks for accurate object detection and segmentation. IEEE Trans Pattern Anal Mach Intell 38(1):142–158CrossRefGoogle Scholar
  12. 12.
    He K, Zhang X, Ren S et al (2015) Spatial pyramid pooling in deep convolutional networks for visual recognition. IEEE Trans Pattern Anal Mach Intell 37 (9):1904–1916.  https://doi.org/10.1109/TPAMI.2015.2389824 CrossRefGoogle Scholar
  13. 13.
    He K, Zhang X, Ren S et al (2016) Deep residual learning for image recognition. In: Computer vision and pattern recognition, pp 770–778. Las Vegas, NV, USAGoogle Scholar
  14. 14.
    Jazayeri A, Cai H, Zheng JY et al (2011) Vehicle detection and tracking in car video based on motion model. IEEE Trans Intell Transp Syst 12(2):583–595CrossRefGoogle Scholar
  15. 15.
    Kong T, Yao A, Chen Y et al (2016) Hypernet: Towards accurate region proposal generation and joint object detection. IEEE Conf Comput Vis Pattern Recognit, pp 845–853Google Scholar
  16. 16.
    Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. Int Conf Neural Inf Process Syst 1:1097–1105Google Scholar
  17. 17.
    Leitloff J, Rosenbaum D, Kurz F et al (2014) An operational system for estimating road traffic information from aerial images. Remote Sens 6(11):11315–11341CrossRefGoogle Scholar
  18. 18.
    Liu W, Anguelov D, Erhan D et al (2016) Ssd: Single shot multibox detector. Eur Conf Comput Vis 9905:21–37CrossRefGoogle Scholar
  19. 19.
    Liu K, Mattyus G (2015) Fast multiclass vehicle detection on aerial images. IEEE Geosci Remote Sens Lett 12(9):1938–1942.  https://doi.org/10.1109/LGRS.2015.2439517 CrossRefGoogle Scholar
  20. 20.
    Lu Y, Chowdhery A, Kandula S (2016) Optasia: a relational platform for efficient large-scale video analytics. ACM Symposium on Cloud Computing, pp 57–70Google Scholar
  21. 21.
    Noh SJ, Shim D, Jeon M (2016) Adaptive sliding-window strategy for vehicle detection in highway environments. IEEE Trans Intell Transp Syst 17(2):323–335CrossRefGoogle Scholar
  22. 22.
    Redmon J, Divvala S, Girshick R et al (2015) You only look once: unified, real-time object detection. IEEE Conf Comput Vis Pattern Recognit, pp 779–788Google Scholar
  23. 23.
    Ren S, He K, Girshick R et al (2015) Faster r-cnn: Towards real-time object detection with region proposal networks. IEEE Trans Pattern Anal Mach Intell 39(6):1137–1149.  https://doi.org/10.1109/TPAMI.2016.2577031 CrossRefGoogle Scholar
  24. 24.
    Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. Comput Sci, pp 1–13Google Scholar
  25. 25.
    Sun Z, Bebis G, Miller R (2002) On-road vehicle detection using gabor filters and support vector machines. Int Conf Digital Signal Process 2:1019–1022Google Scholar
  26. 26.
    Sun Z, Bebis G, Miller R (2005) On-road vehicle detection using evolutionary gabor filter optimization. IEEE Trans Intell Transp Syst 6(2):125–137CrossRefGoogle Scholar
  27. 27.
    Sun Z, Bebis G, Miller R (2006) On-road vehicle detection: a review. IEEE Trans Pattern Anal Mach Intell 28(5):694–711CrossRefGoogle Scholar
  28. 28.
    Szegedy C, Liu W, Jia Y et al (2015) Going deeper with convolutions. In: IEEE Conference on computer vision and pattern recognition, pp 1–9. Boston, MA, USAGoogle Scholar
  29. 29.
    Uijlings JR, Sande KE, Gevers T et al (2013) Selective search for object recognition. Int J Comput Vis 104(2):154–171.  https://doi.org/10.1007/s11263-013-0620-5 CrossRefGoogle Scholar
  30. 30.
    Wang G, Wang X, Fan B et al (2017) Feature extraction by rotation-invariant matrix representation for object detection in aerial image. IEEE Geosci Remote Sens Lett PP(99):1–5Google Scholar
  31. 31.
    Zhang X, Zou J, Ming X et al (2015) Efficient and accurate approximations of nonlinear convolutional networks. In: IEEE conference on computer vision and pattern recognition, pp 1984–1992. Boston, MA, USAGoogle Scholar

Copyright information

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

Authors and Affiliations

  1. 1.College of Information Science and EngineeringHenan University of TechnologyZhengzhouChina

Personalised recommendations