A New Variational Framework for Rigid-Body Alignment
- Cite this paper as:
- Kato T., Tsuda K., Tomii K., Asai K. (2004) A New Variational Framework for Rigid-Body Alignment. In: Fred A., Caelli T.M., Duin R.P.W., Campilho A.C., de Ridder D. (eds) Structural, Syntactic, and Statistical Pattern Recognition. SSPR /SPR 2004. Lecture Notes in Computer Science, vol 3138. Springer, Berlin, Heidelberg
We present a novel algorithm for estimating the rigid-body transformation of a sequence of coordinates, aiming at the application to protein structures. Basically the sequence is modeled as a hidden Markov model where each state outputs an ellipsoidal Gaussian. Since maximum likelihood estimation requires to solve a complicated optimization problem, we introduce a variational estimation technique, which performs singular value decomposition in each step. Our probabilistic algorithm allows to superimpose a number of sequences which are rotated and translated in arbitrary ways.
Unable to display preview. Download preview PDF.