Direct Model-Based Tracking of 3D Object Deformations in Depth and Color Video

被引:0
作者
Andreas Jordt
Reinhard Koch
机构
[1] University of Kiel,Multimedia Information Processing Group
来源
International Journal of Computer Vision | 2013年 / 102卷
关键词
Deformation; Tracking; Range video;
D O I
暂无
中图分类号
学科分类号
摘要
The tracking of deformable objects using video data is a demanding research topic due to the inherent ambiguity problems, which can only be solved using additional assumptions about the deformation. Image feature points, commonly used to approach the deformation problem, only provide sparse information about the scene at hand. In this paper a tracking approach for deformable objects in color and depth video is introduced that does not rely on feature points or optical flow data but employs all the input image information available to find a suitable deformation for the data at hand. A versatile NURBS based deformation space is defined for arbitrary complex triangle meshes, decoupling the object surface complexity from the complexity of the deformation. An efficient optimization scheme is introduced that is able to calculate results in real-time (25 Hz). Extensive synthetic and real data tests of the algorithm and its features show the reliability of this approach.
引用
收藏
页码:239 / 255
页数:16
相关论文
共 43 条
[1]  
Alizadeh F.(2001)Second-order cone programming Mathematical Programming 95 3-51
[2]  
Goldfarb D.(1998)A parametric deformable model to fit unstructured 3d data Computer Vision and Image Understanding 71 39-54
[3]  
Bardinet E.(2010)3d deformable face tracking with a commodity depth camera Camera 6313 229-242
[4]  
Cohen L. D.(1991)Finite element methods for active contour models and balloons for 2d and 3d images IEEE Transactions on Pattern Analysis and Machine Intelligence 15 1131-1147
[5]  
Ayache N.(2006)Non-rigid stereo factorization International Journal of Computer Vision 66 193-207
[6]  
Cai Q.(2007)Non-rigid structure from motion using ranklet-based tracking and non-linear optimization Image and Vision Computing 25 297-310
[7]  
Gallup D.(1991)Rigid body motion from range image sequences CVGIP. Image Understanding 53 1-13
[8]  
Zhang C.(1993)Dynamic 3-d scene analysis through synthesis feedback control IEEE Transactions on Pattern Analysis and Machine Intelligence 15 556-568
[9]  
Zhang Z.(1985)Algorithms for estimation of three-dimensional motion AT & T Bell Laboratories Technical Journal 64 2-49
[10]  
Cohen L. D.(1988)Fronts propagating with curvature dependent speed: algorithms based on Hamilton-Jacobi formulations Journal of Computational Physics 79 12-122