5.3
Pose computation aims to estimate the transformation needed to map an object model from the model coordinate system into the sensory data [Ayache and Faugeras 1986,Faugeras and Hebert 1986,Bolles and Horaud 1986,Stockman 1987,Haralick et al. 1989,Grimson 1990,Umeyama 1991]. In this section, we derive two MRF models for pose estimation. The first is a model for pose clustering from corresponding point data containing multiple poses and outliers. The second model attempts to solve 3D matching and pose simultaneously from a 2D image without using view-invariants.