Comparison of Appearance-Based and Geometry-Based Bubble Detectors

  • Nataliya Strokina
  • Roman Juránek
  • Tuomas Eerola
  • Lasse Lensu
  • Pavel Zemčik
  • Heikki Kälviäinen
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8671)


Bubble detection is a complicated tasks since varying lighting conditions changes considerably the appearance of bubbles in liquid. The two common techniques to detect circular objects such as bubbles, the geometry-based and appearance-based approaches, have their advantages and weaknesses. The geometry-based methods often fail to detect small blob-like bubbles that do not match the used geometrical model, and appearance-based approaches are vulnerable to appearance changes caused by, e.g., illumination. In this paper, we compare a geometry-based concentric circular arrangements (CCA) and appearance-based sliding window methods as well as their combinations in terms of bubble detection, gas volume computation, and size distribution estimation. The best bubble detection performance was achieved with the sliding window method whereas the most precise volume estimate was produced by the CCA method. The combination of the two approaches gave only a minor advantage compared to the base methods.


Ground Truth Bubble Size Bubble Volume Weak Hypothesis Sequential Probability Ratio Test 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Canny, J.: A computational approach to edge detection. PAMI 8(6), 679–698 (1986)CrossRefGoogle Scholar
  2. 2.
    Dence, C.W., Reeve, D.W.: Pulp Bleaching, Principles and Practice. TAPPI (1996)Google Scholar
  3. 3.
    Dominguez, R.A., Corkidi, G.: Automated recognition of oil drops in images of multiphase dispersions via gradient direction pattern. In: CISP, vol. 3, pp. 1209–1213 (2011)Google Scholar
  4. 4.
    Duda, R., Hart, P.: Using the hough transform to detect lines and curves in pictures. Comm ACM, pp. 11–15 (1972)Google Scholar
  5. 5.
    Fischler, M.A., Bolles, R.C.: Random sample consensus: A paradigm for model fitt. with appl. to image anal. and autom. cartogr. Comm ACM 24(6), 381–395 (1981)CrossRefMathSciNetGoogle Scholar
  6. 6.
    Herout, A., Jošth, R., Juránek, R., Havel, J., Hradiš, M., Zemčík, P.: Real-time object detection on cuda. JRIP (2011)Google Scholar
  7. 7.
    Leavers, V.: Which Hough Transform? GMIP: IU 58(2), 250–264 (1993)Google Scholar
  8. 8.
    Press, W., Flannery, B., Teukolsky, S., Vetterling, W.: Numerical Recipes in C: The Art of Scientific Computing. Cambridge University Press (1992)Google Scholar
  9. 9.
    Ronneberger, O., Wang, Q., Burkhardt, H.: Fast and robust segm. of sph.l particles in vol. data sets from brightfield microsc. In: Proc. ISBI, pp. 372–375 (2008)Google Scholar
  10. 10.
    Strokina, N., Matas, J., Eerola, T., Lensu, L., Kälviäinen, H.: Detection of bubbles as concentric circular arrangements. In: ICPR, pp. 2655–2659 (2012)Google Scholar
  11. 11.
    Strokina, N.: Machine vision methods for process measurements in pulping. Ph.D. thesis, LUT (2013)Google Scholar
  12. 12.
    Viola, P., Jones, M.J.: Robust real-time face detection. IJCV (2004)Google Scholar
  13. 13.
    Šochman, J., Matas, J.: Waldboost - learning for time constrained sequential detection. In: CVPR (2005)Google Scholar
  14. 14.
    Zabulis, X., Papara, M., Chatziargyriou, A., Karapantsios, T.D.: Detection of densely dispersed spherical bubbles in dig. images based on a templ. matching technique. appl. to wet foams. Colloids and Surfaces A: PEA 309, 96–106 (2007)CrossRefGoogle Scholar
  15. 15.
    Zemčík, P., Juránek, R., Musil, M., Musil, P., Hradiš, M.: High performance architecture for object detection in streamed videos. In: Proc ICFPLA (2013)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Nataliya Strokina
    • 1
  • Roman Juránek
    • 2
  • Tuomas Eerola
    • 3
  • Lasse Lensu
    • 3
  • Pavel Zemčik
    • 2
  • Heikki Kälviäinen
    • 3
  1. 1.Department of Signal ProcessingTampere University of TechnologyTampereFinland
  2. 2.Department of Computer Graphics and MultimediaBrno University of TechnologyBrnoCzech Republic
  3. 3.Machine Vision and Pattern Recognition LaboratoryLappeenranta University of TechnologyLappeenrantaFinland

Personalised recommendations