Maximum a Posteriori Estimation of Linear Shape Variation With Application to Vertebra and Cartilage Modeling

The estimation of covariance matrices is a crucial step in several statistical tasks. Especially when using few samples of a high dimensional representation of shapes, the standard maximum likelihood estimation (ML) of the covariance matrix can be far from the truth, is often rank deficient, and may lead to unreliable results. In this paper, we discuss regularization by prior knowledge using maximum a posteriori (MAP) estimates. We compare ML to MAP using a number of priors and to Tikhonov regularization. We evaluate the covariance estimates on both synthetic and real data, and we analyze the estimates' influence on a missing-data reconstruction task, where high resolution vertebra and cartilage models are reconstructed from incomplete and lower dimensional representations. Our results demonstrate that our methods outperform the traditional ML method and Tikhonov regularization.

[1]  T. W. Anderson.  An Introduction to Multivariate Statistical Analysis , 1959 .

[2]  P. Thomas Fletcher,et al.  Principal geodesic analysis for the study of nonlinear statistics of shape , 2004, IEEE Transactions on Medical Imaging.

[3]  J. Tukey,et al.  Multiple-Factor Analysis , 1947 .

[4]  J. Magnus,et al.  Matrix Differential Calculus with Applications in Statistics and Econometrics (Revised Edition) , 1999 .

[5]  Marleen de Bruijne,et al.  Quantitative vertebral morphometry using neighbor-conditional shape models , 2007, Medical Image Anal..

[6]  P. Thomas Fletcher,et al.  Multi-scale 3-D Deformable Model Segmentation Based on Medial Description , 2001, IPMI.

[7]  I. Jolliffe.  Principal Component Analysis , 2002 .

[8]  J. Friedman.  Regularized Discriminant Analysis , 1989 .

[9]  T. Spector,et al.  The assessment of vertebral deformity: A method for use in population studies and clinical trials , 1993, Osteoporosis International.

[10]  Kari Karhunen,et al.  Über lineare Methoden in der Wahrscheinlichkeitsrechnung , 1947 .

[11]  B. Efron,et al.  Multivariate Empirical Bayes and Estimation of Covariance Matrices , 1976 .

[12]  J. Magnus,et al.  Matrix Differential Calculus with Applications in Statistics and Econometrics , 1991 .

[13]  P. Thomas Fletcher,et al.  Automatic shape model building based on principal geodesic analysis bootstrapping , 2008, Medical Image Anal..

[14]  K. Strimmer,et al.  Statistical Applications in Genetics and Molecular Biology A Shrinkage Approach to Large-Scale Covariance Matrix Estimation and Implications for Functional Genomics , 2011 .

[15]  L. R. Haff.  Empirical Bayes Estimation of the Multivariate Normal Covariance Matrix , 1980 .

[16]  C. Goodall.  Procrustes methods in the statistical analysis of shape , 1991 .

[17]  Klaus Baggesen Hilger,et al.  Statistical shape analysis using non-Euclidean metrics , 2003, Medical Image Anal..

[18]  Jon Sporring,et al.  Bayes Reconstruction of Missing Teeth , 2008, Journal of Mathematical Imaging and Vision.

[19]  Michael E. Tipping,et al.  Probabilistic Principal Component Analysis , 1999 .

[20]  R. Eastell,et al.  Classification of vertebral fractures , 1991, Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research.

[21]  Thomas Vetter,et al.  Reconstructing the Complete 3D Shape of Faces from Partial Information (Rekonstruktion der dreidimensionalen Form von Gesichtern aus partieller Information) , 2002, Informationstechnik Tech. Inform..

[22]  H. Hotelling.  Analysis of a complex of statistical variables into principal components. , 1933 .

[23]  Karl Pearson F.R.S..  LIII. On lines and planes of closest fit to systems of points in space , 1901 .

[24]  Marleen de Bruijne,et al.  Semiautomatic segmentation of vertebrae in lateral x-rays using a conditional shape model. , 2007, Academic radiology.

[25]  Timothy F. Cootes,et al.  Use of active shape models for locating structures in medical images , 1994, Image Vis. Comput..

[26]  P. Switzer,et al.  A transformation for ordering multispectral data in terms of image quality with implications for noise removal , 1988 .

[27]  Y. Amit,et al.  Towards a coherent statistical framework for dense deformable template estimation , 2007 .

[28]  C. Anderson‐Cook,et al.  An Introduction to Multivariate Statistical Analysis (3rd ed.) (Book) , 2004 .

[29]  F. Bookstein.  Shape and the information in medical images , 1996 .

[30]  C. Stein,et al.  Estimation with Quadratic Loss , 1992 .

[31]  H. K. Genant,et al.  Vertebral fracture risk (VFR) score for fracture prediction in postmenopausal women , 2011, Osteoporosis International.

[32]  Marleen de Bruijne,et al.  Quantitative Vertebral Morphometry Using Neighbor-Conditional Shape Models , 2006, MICCAI.

[33]  A. E. Hoerl,et al.  Ridge regression: biased estimation for nonorthogonal problems , 2000 .

[34]  J. Neyman,et al.  INADMISSIBILITY OF THE USUAL ESTIMATOR FOR THE MEAN OF A MULTIVARIATE NORMAL DISTRIBUTION , 2005 .

[35]  Seymour Geisser,et al.  Discrimination, Allocatory and Separatory, Linear Aspects , 1977 .

[36]  P. Thomas Fletcher,et al.  Gaussian Distributions on Lie Groups and Their Application to Statistical Shape Analysis , 2003, IPMI.

[37]  Nicholas Ayache,et al.  Clinical DT-MRI estimation, smoothing and fiber tracking with Log-Euclidean metrics , 2006, 3rd IEEE International Symposium on Biomedical Imaging: Nano to Macro, 2006..

[38]  Timothy F. Cootes,et al.  Active Shape Models-Their Training and Application , 1995, Comput. Vis. Image Underst..