Advertisement

Adaptive Learning for Correlation Filter Object Tracking

  • Dongcheng ChenEmail author
  • Jingying Hu
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 905)

Abstract

To solve the real-time quality and adaptive quality of traditional correlation algorithm. An adaptive correlation filter tracking algorithm is proposed. First, training the filter using machine learning, make the algorithm be adaptive to the object changing. Then, weighting the image patch with a cosine window, the object region has larger weighting value than the edge region, which ensures the continuity of the Cyclic matrix. At last calculate the response matrix using the convolution of the input image patch and the filter matrix in Fourier domain. Experiment on various videos shows that for 26 pixel × 24 pixel object, filter the image in a 130 × 120 pixel patch, the processing speed could be 210 fps. The proposed tracking algorithm can track object with good timing quality and robustly.

Keywords

Correlation filter Weighted Fourier domain Cyclic matrix 

References

  1. 1.
    Zhang, T., Xu, C., Yang, M.H.: Multi-task correlation particle filter for robust object tracking. In: IEEE Conference on Computer Vision and Pattern Recognition, pp. 4819–4827. IEEE Computer Society (2017)Google Scholar
  2. 2.
    Yang, Y., Bilodeau, G.A.: Multiple object tracking with kernelized correlation filters in urban mixed traffic. In: IEEE Computer Society, pp. 209–216 (2017)Google Scholar
  3. 3.
    Koubaa, A., Qureshi, B.: DroneTrack: cloud-based real-time object tracking using unmanned aerial vehicles over the internet. IEEE Access 6(99), 13810–13824 (2018)CrossRefGoogle Scholar
  4. 4.
    Duda, R., Hart, P.: Pattern Classification and Scene Analysis, pp. 276–284. Wiley, New York (1973)zbMATHGoogle Scholar
  5. 5.
    Sharma, S., Ansari, J.A., Murthy, J.K., Krishna, M.: Beyond pixels: leveraging geometry and shape cues for online multi-object tracking. In: IEEE International Conference on Robotics and Automation. IEEE (2018)Google Scholar
  6. 6.
    Hester, C.F., Casasent, D.: Multivariant technique for multiclass pattern recognintion. Appl. Opt. 19(11), 1758–1761 (1980)CrossRefGoogle Scholar
  7. 7.
    Vijaya Kumar, B.V.K.: Minmum-variance synthetic discriminant functions. Opt. Eng. 31, 915 (1992)CrossRefGoogle Scholar
  8. 8.
    Mahalanobis, A., Vijaya Kumar, B.V.K., Casasent, D.: Minimum average correlation energy filters. Appl. Opt. 26(17), 3633 (1987)CrossRefGoogle Scholar
  9. 9.
    Bolme, D.S., Draper, B.A., Beveridge, J.R.: Average of synthetic exact filters. In: IEEE Conference on Computer Vision and Pattern Recognition (2009)Google Scholar
  10. 10.
    Bolme, D.S., Beveridge, J.R., Draper, B.A., Lui, Y.M.: Visual object tracking using adaptive correlation filters. In: IEEE Conference on Computer Vision and Pattern Recognition (2010)Google Scholar
  11. 11.
    Henriques, J.F., Caseiro, R., Martins, P., et al.: Exploiting the circulant structure of tracking-by-detection with kernels. In: Europe Conferance on Computer Vision (2014)Google Scholar
  12. 12.
    Hon, S., Wathen, A.: Circulant preconditioners for analytic functions of Toeplitz matrices. Numer. Algorithms (suppl. C), 1–20 (2018) Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Jiangsu Automation Research InstituteLianyungangChina

Personalised recommendations