Structure from Motion Based Approaches to 3D Reconstruction in Minimal Invasive Laparoscopy

  • Andrés F. Mármol Vélez
  • Jan Marek Marcinczak
  • Rolf-Rainer Grigat
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7325)


The present article proposes a Structure from Motion (SfM) methodology to recover the liver surface from endoscopic video sequences. Features from an imaged liver are extracted and tracked for the complete sequence to generate a correspondences lookup table (C-LUT) between all frames. A keyframe selection code extracts two frames, from which the relative pose of the camera is reconstructed using a MSAC-based 5-Point algorithm. Techniques such as an optimal triangulation method and a PnP resection algorithm are also used to obtain an initial 3D surface of the liver. A global Bundle Adjustment step refines the initial reconstruction. Proper parametrization and conditioning of these techniques are compared and evaluated under typical laparoscopic uncertainties arising from patient, illumination, reflections, image quality and organs’ location among others. A robotic system and grid patterns are used to provide camera pose and surface ground truth data respectively.


Laparoscopy Endoscopy Structure from Motion Image Reconstruction Computer Aided Diagnosis Epipolar geometry Optimal triangulation PnP problem Bundle Adjustment Rigid Registration 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Jones, D., Wu, J., Soper, N.: Laparoscopic surgery: principles and procedures. Marcel Dekker Inc., New York (2004)Google Scholar
  2. 2.
    Pollefeys, M., van Gool, L., Vergauwen, M.: Visual modeling with a hand-held camera. Int. J. Comput. Vision 59, 207–232 (2004)CrossRefGoogle Scholar
  3. 3.
    Stoyanov, D., Mylonas, G.P., Deligianni, F., Darzi, A., Yang, G.-Z.: Soft-Tissue Motion Tracking and Structure Estimation for Robotic Assisted MIS Procedures. In: Duncan, J.S., Gerig, G. (eds.) MICCAI 2005. LNCS, vol. 3750, pp. 139–146. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  4. 4.
    Mountney, P., Stoyanov, D., Yang, G.Z.: Three-dimensional tissue deformation recovery and tracking. Signal Processing Magazine 27, 14–24 (2010)CrossRefGoogle Scholar
  5. 5.
    Lowe, D.: Distinctive image features from scale-invariant keypoints. Int. J. Comput. Vision 60, 91–110 (2004)CrossRefGoogle Scholar
  6. 6.
    Nistér, D.: An efficient solution to the five-point relative pose problem. IEEE Trans. Pattern Anal. Mach. Intell. 26, 756–777 (2004)CrossRefGoogle Scholar
  7. 7.
    Lepetit, V., Moreno-Noguer, F., Fua, P.: EPnP: An accurate O(n) solution to the PnP problem. Int. J. Comput. Vision 81, 155–166 (2009)CrossRefGoogle Scholar
  8. 8.
    Fischler, M., Bolles, R.: Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography. J. Communications ACM 24, 381–395 (1981)MathSciNetCrossRefGoogle Scholar
  9. 9.
    Hartley, R., Zisserman, A.: Multiple view geometry. Cambridge University Press, Cambridge (2004)zbMATHCrossRefGoogle Scholar
  10. 10.
    Rabaud, V.: Structure from motion toolbox,
  11. 11.
    von Oehsen, U., Marcinczak, J.M., Mármol Vélez, A., Grigat, R.-R.: Key-frame selection for robust pose estimation in laparoscopic videos. In: Medical Imaging 2012: Image-Guided Procedures, Robotic Interventions, and Modeling, New York (2012)Google Scholar
  12. 12.
    Bern, M., Eppstein, D., Gilbert, J.: Provably Good Mesh Generation. J. Comput. Syst. Sci. 48, 231–241 (1990)Google Scholar
  13. 13.
    Camera calibration toolbox for Matlab,
  14. 14.
    Myronenko, A., Song, X.: Point-Set Registration: Coherent Point Drift. IEEE Trans. on Pattern Analysis and Machine Intelligence 32, 2262–2275 (2010)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Andrés F. Mármol Vélez
    • 1
  • Jan Marek Marcinczak
    • 1
  • Rolf-Rainer Grigat
    • 1
  1. 1.Vision SystemsHamburg University of TechnologyHamburgGermany

Personalised recommendations