Department of Computer Science Technische Universitaet Muenchen
Home People Teaching Research Publications Info Sitemap
People | Teaching | Research | Publications | Job offers

Publications

Sorted by DateClassified by Publication TypeClassified by Research CategoryClassified by Research GroupClassified by Author Last Name

Matthias Wimmer and Bernd Radig. Initial Pose Estimation for 3D Models Using Learned Objective Functions. In Proceedings of the 8th Asian Conference on Computer Vision (ACCV07), pp. 332–341, Springer, Tokyo, Japan, November 2007.

Download

[PDF]13.9MB  

Abstract

Tracking 3D models in image sequences essentially requires determining their initial position and orientation. Our previous work identifies the objective function as a crucial component for fitting 2D models to images. We state preferable properties of these functions and we propose to learn such a function from annotated example images. This paper extends this approach by making it appropriate to also fit 3D models to images. The correctly fitted model represents the initial pose for model tracking. However, this extension induces nontrivial challenges such as out-of-plane rotations and self occlusion, which cause large variation to the models surface visible in the image. We solve this issue by connecting the input features of the objective function directly to the model. Furthermore, sequentially executing objective functions specifically learned for different displacements from the correct positions yields highly accurate objective values.

BibTeX Entry

@inproceedings{ wimmer07initial,
  author    		= {Matthias Wimmer and Bernd Radig},
  title     		= {Initial Pose Estimation for {3D}~Models Using Learned Objective Functions},
  booktitle 		= {Proceedings of the 8th~Asian Conference on Computer Vision~(ACCV07)},
  origbooktitle 		= {Computer Vision -- ACCV 2007},
  pages     		= {332-341},
  editor    		= {Yasushi Yagi and Sing Bing Kang and In So Kweon and Hongbin Zha},
  year      		= {2007},
  month				= {November},
  address   		= {Tokyo, Japan},
  publisher 		= {Springer},
  location  		= {Heidelberg},
  series    		= {LNCS},
  volume    		= {4844},
  isbn      		= {978-3-540-76389-5},
  bib2html_pubtype	= {Refereed Conference Paper},
  bib2html_rescat   = {Image Understanding},
  bib2html_groups   = {IU},
  abstract       	= 
  						{Tracking 3D models in image sequences essentially requires determining their initial position and orientation. Our previous work identifies the objective function as a crucial component for fitting 2D models to images. We state preferable properties of these functions and we propose to learn such a function from annotated example images.
						This paper extends this approach by making it appropriate to also fit 3D models to images. The correctly fitted model represents the initial pose for model tracking. However, this extension induces nontrivial challenges such as out-of-plane rotations and self occlusion, which cause large variation to the models surface visible in the image.
						We solve this issue by connecting the input features of the objective function directly to the model. Furthermore, sequentially executing objective functions specifically learned for different displacements from the correct positions yields highly accurate objective values.},
}

Generated by bib2html.pl (written by Patrick Riley ) on Thu Aug 28, 2008 17:40:35 UTC


Computer Science IX | Technische Universität München