Dense 3D Reconstruction and Tracking of Dynamic Surface - Advances in Image and Graphics Technologies - page 222

4 Conclusions

In this essay, we have presented a framework to reconstruct and track the dynamic 3D
surface from stereo video by optimizing an energy function. Our approach relies on
finding the correspondence among four blocks in a stereo image sequence by
obtaining the parameters of transformation, and after all block correspondences are
found, all blocks can be integrated into a dense 3D surface by bilinear interpolation.
The simulated and real data experiments confirm the performance of our approach.
The dense motion field of the 3D surface (which provides an approach to analyze the
motion rules of dynamic surface) can be achieved. Since the different block
correspondences in one surface can be found independently, it is possible to
parallelize block matching in a shared memory environment to speed up the process
of reconstruction and tracking. In our future work, we intend to find out the solution
to the self-occlusion problem in our framework which is not considered in this essay.

References

1. Ahmed, N., Theobalt, C., R ”ossl, C., Thrun, S., Seidel, H.: Dense Correspondence
Finding for Parametrization-free Animation Reconstruction from Video. In: Proceedings
of Computer Vision and Pattern Recognition (2008)
2. Baker, S., Matthews, I.: Lucas-kanade 20 years on: A unifying framework. International
Journal of Computer Vision 56(3), 221-255 (2004)
3. Ballan, L., Cortelazzo, G.: Marker-less motion capture of skinned models in a four camera
set-up using optical flow and silhouettes. In: Int. Symp. on 3DPVT (2008)
4. Chai, M., Wang, L., Weng, Y., Jin, X., Zhou, K.: Dynamic hair manipulation in images
and videos. To appear in ACM TOG 32, 4 (2013)
5. De Aguiar, E., Theobalt, C., Stoll, C., Seidel, H.: Marker-less deformable mesh tracking
for human shape and motion capture. In: Proc. CVPR (2007)
6. Doshi, A., Hilton, A., Starck, J.: An empirical study of non-rigid surface feature matching.
In: Visual Media Production (CVMP 2008), 5th European Conference on. pp. 1-10 (2008)
7. Furukawa, Y., Ponce, J.: Dense 3d motion capture for human faces. In: Proc. CVPR (2009)
8. Groeger, M., Ortmaier, T., Sepp, W., Hirzinger, G.: Tracking local motion on the beating
heart. In: Proceedings of SPIE. vol. 4681, p. 233 (2002)
9. Hilsmann, A., Eisert, P.: Tracking deformable surfaces with optical flow in the presence of
self occlusions in monocular image sequences. In: CVPR Workshop on Non-Rigid Shape
Analysis and Deformable Image Alignment, Anchorage, USA (2008)
10. Huguet, F., Devernay, F.: A variational method for scene flow estimation from stereo
sequences. Research Report 6267
11. Noce, A., Triboulet, J., Poignet, P., CNRS, M.: Efficient tracking of the heart using
texture. In: Engineering in Medicine and Biology Society, 2007. EMBS 2007. 29th Annual
International Conference of the IEEE. pp. 4480-4483 (2007)
12. Pekelny, Y., Gotsman, C.: Articulated object reconstruction and markerless motion capture
from depth video. In: Computer Graphics Forum. vol. 27, pp. 399-408. Citeseer (2008)
13. Richa, R., Poignet, P., Liu, C.: Deformable motion tracking of the heart surface. In:
IEEE/RSJ International Conference on Intelligent Robots and Systems, 2008. IROS 2008.
pp. 3997-4003 (2008)