GPU Based Horn-Schunck Method to Estimate Optical Flow and Occlusion

  • Vanel LazcanoEmail author
  • Francisco Rivera
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11436)


Optical flow is the apparent motion pattern of pixels in two consecutive images. Optical flow has many applications: navigation control of autonomous vehicles, video compression, noise suppression, and others. There are different methods to estimate the optical flow, where variational models are the most frequently used. These models state an energy model to compute the optical flow. These models may fail in presence of occlusions and illumination changes. In this work is presented a method that estimates the flow from the classical Horn-Schunk method and the incorporation of an occlusion layer that gives to the model the ability to handle occlusions. The proposed model was implemented in an Intel i7, 3.5 GHz, GPU GeForce NVIDIA-GTX-980-Ti, using a standard webcam. Using images of \(320\times 240\) pixels we reached 4 images per second, i.e. this implementation can be used in an application like an autonomous vehicle.


Optical flow Occlusion estimation Variational model 


  1. 1.
    Baker, S., Scharstein, D., Lewis, J., Roth, S., Black, M., Szelensky, R.: A database and evaluation methodology for optical flow. Int. J. Comput. Vis. 92, 1–31 (2011)CrossRefGoogle Scholar
  2. 2.
    Ballester, C., Garrido, L., Lazcano, V., Caselles, V.: A TV-L1 optical flow method with occlusion detection. In: Pinz, A., Pock, T., Bischof, H., Leberl, F. (eds.) DAGM/OAGM 2012. LNCS, vol. 7476, pp. 31–40. Springer, Heidelberg (2012). Scholar
  3. 3.
    Barron, J., Fleet, D., Beauchemin, S.: Performance of optical flow technics. Int. J. Comput. Vis. 12(1), 43–47 (2011)CrossRefGoogle Scholar
  4. 4.
    Garamendi, J.F., Ballester, C., Garrido, L., Lazcano, V.: Joint TV-L1 optical flow and occlusion estimation. IPOL J. - Image Process. On Line, preprint (2014)Google Scholar
  5. 5.
    Horn, B., Schunck, B.G.: Determining optical flow. Artif. Intell. 17, 185–204 (1981)CrossRefGoogle Scholar
  6. 6.
    Lazcano, V., Garrido, L., Ballester, C.: Jointly optical flow and occlusion estimation for images with large displacements. In: SciTePress (ed.) Proceedings of the 13th International Joint Conference on Computer Vision Imaging and Computer Graphics Theory and Applications, pp. 588–595. INSTICC (2018)Google Scholar
  7. 7.
    Sand, P., Teller, S.: Particle video: long-range motion estimation using point trajectory. Int. J. Comput. Vis. 80(1), 72 (2008)CrossRefGoogle Scholar
  8. 8.
    Smirnov, M.: Optical flow estimation with CUDA. NVIDIA white papers (2012)Google Scholar
  9. 9.
    Xiao, J., Cheng, H., Sawhney, H., Rao, C., Isnardi, M.: Bilateral filtering-based optical flow estimation with occlusion detection. In: Leonardis, A., Bischof, H., Pinz, A. (eds.) ECCV 2006. LNCS, vol. 3951, pp. 211–224. Springer, Heidelberg (2006). Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Núcleo de Matemática, Física y EstadísticaUniversidad MayorProvidenciaChile
  2. 2.Universidad MayorProvidenciaChile

Personalised recommendations