Abstract
Viewer-centered estimation of the pose of a three dimensional object has two main advantages: No explicit models are needed and error-prone corner detection is not necessary. Eigenspace methods have been successful in pose estimation especially for faces. However, most eigenspace-based algorithms fail if the images are corrupted, e. g. if the object is occluded, the background differs from the training images or the image is geometrically transformed. EigenTracking by Black and Jepson uses robust estimation to find the correct pose. We show that performance degrades for objects whose silhouette changes greatly with 3D rotation. To solve this problem we introduce masks that adapt to the estimated object pose. To this end we used hierarchical eigenspaces containing both the appearance and mask descriptions. We illustrate the improvement in pose estimation precision for some typical objects.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Pentland, A., Moghaddam, B., Starner, T.: View-based and modular eigenspaces for face recognition. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR 1994), Seattle, USA, vol. 1, pp. 84–91 (1994)
Yoshimura, S., Kanade, T.: Fast template matching based on the normalized correlation by using multiresolution eigenimages. In: IEEE/RSJ/GI International Conference on Intelligent Robots and Systems, Advanced Robotic Systems and the Real World, New York, USA, vol. 3, pp. 2086–2093 (1994)
Chang, C.Y., Maciejewski, A.A., Balakrishnan, V., Roberts, R.G.: Eigendecomposition- based pose detection in the presence of occlusion. In: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Maui, HI, pp. 569–576 (2001)
Ohba, K., Ikeuchi, K.: Detectability, uniqueness and reliability of eigen windows for stable verifikation of partially occluded objects. IEEE Transactions on Pattern Analysis and Machine Intelligence 19, 1043–1048 (1997)
Leonardis, A., Bischof, H.: Robust recognition using eigenimages. Computer Vision and Image Understanding: CVIU 78, 99–118 (2000)
Wildenauer, H., Bischof, H., Leonardis, A.: Eigenspace pyramids for robust and efficient recognition of scaled eigenimages. In: Proc. Computer Vision WinterWorkshop (CVWW), Valtice, Czech Republic, pp. 115–120 (2003)
Black, M.J., Jepson, A.D.: EigenTracking: Robust matching and tracking of articulated objects using a view-based representation. International Journal of Computer Vision 26, 63–84 (1998)
Nene, S.A., Nayar, S.K., Murase, H.: Columbia object image library (COIL-100). Technical Report CUCS-006-96, Columbia University (1996)
Cootes, T.F., Edwards, G.J., Taylor, C.J.: Active appearance models. In: Burkhardt, H., Neumann, B. (eds.) Proc. European Conference on Computer Vision, vol. 2, pp. 484–498. Springer, Heidelberg (1998)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Dörfler, P., Schnurr, C. (2004). Robust Pose Estimation for Arbitrary Objects in Complex Scenes. In: Rasmussen, C.E., Bülthoff, H.H., Schölkopf, B., Giese, M.A. (eds) Pattern Recognition. DAGM 2004. Lecture Notes in Computer Science, vol 3175. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-28649-3_56
Download citation
DOI: https://doi.org/10.1007/978-3-540-28649-3_56
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-22945-2
Online ISBN: 978-3-540-28649-3
eBook Packages: Springer Book Archive