| The pose of an object in three-dimensional space consists of its position and orientation. Pose analysis is concerned with performing computations on functions of pose. This dissertation is concerned with pose analysis in two particular research areas: image registration and protein statistics. Hence, this thesis is divided into two parts. The first part is concerned with problems in image registration and reconstruction of spatial scenes from stereo images. We present several numerical algorithms for six-degree-of-freedom rigid body-registration of fiducial objects to their marks in cross-sectional planar images, given the correspondence between the marks and line fiducials. We also present a method to establish the correspondence between edge features in stereo images using the Hungarian algorithm for optimal assignment. The second part addresses problems in the statistical analysis of protein data. We report patterns in the relative pose between alpha-carbons in proteins and examine the pose statistics of the terminal end of side chains in a binding site relative to an adenosine molecule. We also address the idea of extracting the stiffness and equilibrium distance between residue pair types from distance histograms. |
