New Trends in Surface Reconstruction Using Space-Time Cameras |
57 |
|
|
Figure 34. Reconstruction of Head by perpendicular double stereo rigs (non-circular motion).
Conclusions
In this chapter, an efficient method has been presented to reconstruct the three dimensional model of a moving object by extracting the space curves and tracking
58 |
Hossein Ebrahimnezhad and Hassan Ghassemian |
|
|
them across time using perpendicular double stereo rigs. A new method of space curve extraction on the surface of the object was presented by checking the consistency of torsion and curvature through motion. The nature of space curves makes the extraction very robust against the poor color adjustment between cameras and changing the light angle during the object motion. Projection of the space curves in the camera images were employed for tracking the curves and for the robust motion estimation of the object. The concept of virtual cameras, which are constructed from motion information, was introduced. Constructing the virtual cameras makes possible to use a large number of silhouette cones in the structure- from-silhouette method. Experimental results show that the presented edge based method can be effectively used in reconstruction of visual hull for poorly textured objects. In addition, it does not require the accurate color adjustment during camera setup and provides a better result compared to the other methods, which use the color consistency property. Moreover, the presented method is not limited to the circular turntable rotation. Finally, the double perpendicular stereo setup has been presented as a way to reduce the effect of statistical bias in motion estimation and to enhance the quality of 3D reconstruction. Quantitatively, the average of abstract error in φtotal is reduced from 0.51 deg in single stereo rig to 0.14 deg in perpendicular double stereo rigs and, the average of abstract error in
Ttotal is reduced from 0.23 mm to 0.18 mm for Cow sequence. The respective
values in Helga sequence are φtotal 0.61 - 0.28 deg and |
Ttotal |
0.24-0.20 mm. |
These values are increased in noisy sequence of Helga to |
φtotal |
1.21 - 0.64 deg |
and Ttotal 0.34-0.25 mm. |
|
|
Acknowledgment
This research was supported in part by ITRC, the Iran Telecommunication Research Center, under grant no. TMU 85-05-33. Main part of this chapter has been published previously in [43] by the authors.
References
[1]Y. Han, Geometric algorithms for least squares estimation of 3-D information from monocular image, IEEE Trans. Circuits and Systems for Video Technology, (15) (2) (2005), pp.269-282.
New Trends in Surface Reconstruction Using Space-Time Cameras |
59 |
|
|
[2]T. Papadimitriou, K.I. Diamantaras, M.G. Strintzis, and M. Roumeliotis, Robust Estimation of Rigid-Body 3-D Motion Parameters Based on Point Correspondences, IEEE Trans. Circuits and Systems for Video Technology,
(10) (4) (2000), pp.541-549.
[3]M. Lhuillier, L. Quan, A Quasi-Dense Approach to Surface Reconstruction from Uncalibrated Images, IEEE Trans. Pattern Analysis and Machine Intelligence, (27) (3) (2005), pp. 418-433.
[4]Y. Zhang and C. Kambhamettu, On 3-D Scene Flow and Structure Recovery From Multiview Image Sequences, IEEE Trans. Systems, Man, and Cybernetics - Part B, (33) (4) (2003), pp. 592-606.
[5]P. Eisert, E. Steinbach, and B. Girod, Automatic Reconstruction of Stationary 3-D Objects from Multiple Uncalibrated Camera Views, IEEE Trans. Circuits and Systems for Video Technology, (10) (2) (2000), pp.261277.
[6]Y. Furukawa, A. Sethi, J. Ponce, and D.J. Kriegman, Robust Structure and Motion from Outlines of Smooth Curved Surfaces, IEEE Trans. Pattern Analysis and Machine Intelligence, (28) (2) (2006), pp. 302-315.
[7]B. Triggs, P. McLauchlan, R. Hartley and A. Fiztgibbon, Bundle adjustment- a modern synthesis, In vision Algorithms: Theory and Practice, Lecture Notes on Computer Science, Sppringer-Verlag, (1883) (2000), pp 298-372.
[8]R.I. Hartley and A. Zisserman, Multiple View Geometry in Computer Vision. Cambridge Univ. Press, Second Edition, 2003.
[9]M. Pollefeys, R. Koch, M. Vergauwen, and L. Van Gool, Metric 3D Surface Reconstruction from Uncalibrated Image Sequences, Proc. European Workshop 3D Structure from Multiple Images of Large-Scale Environments,
(1998), pp. 139-154.
[10]M.A. Fischler and R.C. Bolles, Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography, Comm. ACM, (24) (6) (1981), pp. 381-395.
[11]P. Torr and D. Murray, The Development and Comparison of Robust Methods for Estimating the Fundamental Matrix, Int’l J. Computer Vision,
(24) (3) (1997), pp. 271-300.
[12]Z. Zhang, Determining the epiipolar Geometry and its Uncertainity: a Review, Int’l J. Computer Vision, (27) (2) (1998), pp. 161-195.
[13]P. Torr, A. Zisserman, MLESAC: A New robust estimator with application to estimating image geometry, Computer Vision and Image Understanding,
(78) (2000), pp. 138-156.
60 |
Hossein Ebrahimnezhad and Hassan Ghassemian |
|
|
[14]A. Calway, Recursive Estimation of 3D Motion and Surface Structure from Local Affine Flow Parameters, IEEE Trans. Pattern Analysis and Machine Intelligence, (27) (4) (2005), pp. 562-574.
[15]K. Kanatani, 3D interpretation of optical flow by renormalization, Int’l J. Computer Vision, (11) (3) (1993), pp. 267–282.
[16]A. D. Jepson and D. J. Heeger, Linear subspace methods for recovering translation direction, in Spatial Vision in Humans and Robots, Cambridge Univ. Press, (1993), pp. 39–62.
[17]A.K.R. Chowdhury and R. Chellappa, Statistical Bias in 3-D Reconstruction from Monocular Video, IEEE Trans. Image Processing, (14) (8) (2004), pp. 1057-1062.
[18]R. Szeliski and S. B. Kang, Recovering 3D shape and motion from image streams using nonlinear least squares, J. Visual Commun. Image Representation, (5) (1) (1994), pp. 10–28.
[19]G. Young and R. Chellappa, 3-D motion estimation using a sequence of noisy stereo images: Models, estimation, and uniqueness results, IEEE Trans. Pattern Anal. Machine Intell., (12) (1990), pp. 735–759.
[20]J. Weng, P. Cohen, and N. Rebibo, "Motion and Structure Estimation from Stereo Image Sequences ," IEEE Trans. Robotics and Automation, (8) (3) (1992), pp. 362-382.
[21]L. Li and J. Duncan, 3-D translational motion and structure from binocular image flows, IEEE Trans. Pattern Anal. Machine Intell., (15) (1993), pp. 657–667.
[22]F. Dornaika and R. Chung, Stereo correspondence from motion correspondence, Proc. IEEE Conf. Computer Vision Pattern Recognition,
(1999), pp. 70–75.
[23]P.K. Ho and R. Chung, Stereo-Motion with Stereo and Motion in Complement, IEEE Trans. Pattern Anal. Machine Intell., (22) (2) (2000),
pp.215–220.
[24]S.K. Park and I.S. Kweon, Robust and direct estimation of 3-D motion and Scene depth from Stereo image sequences, Pattern Recognition, (34) (9) (2001), pp. 1713-1728.
[25]H. Ebrahimnezhad, H. Ghassemian, 3D Shape Reconstruction of Moving Object by Tracking the Sparse Singular Points, IEEE International workshop on Multimedia Signal Processing, (2006), pp. 192-197.
[26]L. Robert and O.D. Faugeras, Curve-based stereo: figural continuity and curvature, Proc. IEEE Conf. Computer Vision Pattern Recognition, (1991),
pp.57–62.
New Trends in Surface Reconstruction Using Space-Time Cameras |
61 |
|
|
[27]C. Schmid and A. Zisserman, The geometry and matching of lines and curves over multiple views, Int. Journal ComputerVision, (40) (3) (2000),
pp.199–233.
[28]J.H. Han, J. S. Park, Contour Matching Using Epipolar Geometry, IEEE Trans. Pattern Analysis and Machine Intelligence, (22) (4) (2000), pp.358370.
[29]F. Kahl, J. Augut, Multiview Reconstruction of Space Curves, In Int.Conf. Computer Vision, (2003), pp. 181–186.
[30]A.Gray, Modern Differential Geometry of Curves and Surfaces with Mathematica, CRC Press, Second Edition, (1997), pp. 219-222.
[31]A. Bottino, A. Laurentini, Introducing a New Problem: Shape-from- Silhouette When the Relative Positions of the Viewpoints is Unknown,
IEEE Trans. Pattern Analysis and Machine Intelligence, (25) (11) (2003),
pp.1484-1492.
[32]Yang Liu, George Chen, Nelson Max, Christian Hofsetz, Peter McGuinness, Visual Hull Rendering with Multi-view Stereo, Journal of WSCG. (12) (1-3) (2004).
[33]A.Y. Mulayim, U. Yilmaz, V. Atalay, Silhouette-based 3-D model reconstruction from multiple images, IEEE Trans. on Systems, Man and Cybernetics, Part B, (33) (4) (2003), pp. 582-591.
[34]G.K.M. Cheung, S. Baker, T. Kanade, Shape-from-Silhouette Across Time Part I: Theory and Algorithms, Int’l J. Computer Vision, (62) (3) (2005), pp. 221-247.
[35]M. Potmesil, "Generating Octree Models of 3D Objects from their Silhouettes in a Sequence of Images", CVGIP 40, pp. 1-29, 1987.
[36]R.Szeliski, "Rapid Octree Construction from Image Sequences", CVGIP: Image Understanding 58, 1, pp. 2332, July 1993.
[37]W. Matusik, C. Buehler, R. Raskar, S. Gortler, "ImageBased Visual Hulls", SIGGRAPH 2000, McMillan, pp. 369374, July 2328, 2000.
[38]S. Moezzi, A. Katkere, D. Y. Kuramura, and R. Jain, "Reality modeling and visualization from multiple video sequences", IEEE Computer Graphics and Applications, 16( 6), pp. 58– 63, November 1996.
[39]G.K.M. Cheung, T. Kanade, J.Y. Bouguet, and M. Holler. "A real time system for robust 3D voxel reconstruction of human motions", In
Proceedings of the 2000 IEEE Conference on Computer Vision and Pattern Recognition (CVPR ’00), Vol. 2, pp. 714–720, June 2000.
[40]G.K.M. Cheung, et all. "Visual hull alignment and refinement across time: a 3D reconstruction algorithm combining shape-frame-silhouette with stereo."
62 |
Hossein Ebrahimnezhad and Hassan Ghassemian |
|
|
In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition 2003, Vol. 2, pp. 375-382, June 2003.
[41]W. Matusik, et all, "Polyhedral visual hulls for real-time rendering", In
Proceedings of 12th Euro graphics Workshop on Rendering, pp. 115–125, June 2001.
[42]S. Lazebnik, E. Boyer, and J. Ponce. "On computing exact visual hulls of solids bounded by smooth surfaces", In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR'01), Kauai HI, December 2001.
[43]H. Ebrahimnezhad, H. Ghassemian, “Robust Motion from Space Curves and 3D Reconstruction from Multiviews Using Perpendicular Double Stereo Rigs, Journal of Image and Vision Computing, Elsevier,Vol 26, No.10, pp. 1397-1420, Oct. 2008.