Presentation 2015-09-08
Model Selection from the q-Exponential Distribution for Constructing an Organ Point Distribution Model
Mitsunori Yamada, Hidekata Hontani, Hiroshi Matsuzoe,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The authors propose a method that selects an appropriate distribution model for accurately representing statistics of the shape of an organ with a PDM (Point Distribution Model). Normal distributions are widely employed for representing the statistical models in PDMs because the normal distributions have small number of parameters to be estimated and because the normal distributions make the inference for the registration of the models easier. A set of training data often includes some data of which Mahalanobis distances from the mean are much larger than those of the other data and the values of the estimated parameters can be inaccurate when such the data are included if you employ the normal distribution for the model representation. The authors employ a $q$-exponential distribution in order to handle this problem. The $q$-exponential distribution includes the normal distribution and the student-t distribution. Given a set of training data, the proposed method selects an appropriate distribution from the $q$-exponential ones for each region of a target organ in order to represent the statistical model of the target organ accurately. Using the proposed method for constructing a statistical model of the liver, the authors found that distributions with longer tails were selected in the regions that faced the inside of the body.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Point Distribution Model(PDM) / Model Selection / q-exponential distribution / AIC
Paper # MI2015-54
Date of Issue 2015-09-01 (MI)

Conference Information
Committee MI
Conference Date 2015/9/8(1days)
Place (in Japanese) (See Japanese page)
Place (in English) Univ. of Electro-communications
Topics (in Japanese) (See Japanese page)
Topics (in English) Machine learning and its application to medical image processing
Chair Yoshitaka Masutani(Hiroshima City Univ.)
Vice Chair Yoshiki Kawata(Tokushima Univ.) / Yuichi Kimura(Kinki Univ.)
Secretary Yoshiki Kawata(Aichi Inst. of Tech.) / Yuichi Kimura(Nagoya Inst. of Tech.)
Assistant Ryo Haraguchi(NCVC) / Yasushi Hirano(Yamaguchi Univ.)

Paper Information
Registration To Technical Committee on Medical Imaging
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Model Selection from the q-Exponential Distribution for Constructing an Organ Point Distribution Model
Sub Title (in English)
Keyword(1) Point Distribution Model(PDM)
Keyword(2) Model Selection
Keyword(3) q-exponential distribution
Keyword(4) AIC
1st Author's Name Mitsunori Yamada
1st Author's Affiliation Nagoya Institute of Technology(NIT)
2nd Author's Name Hidekata Hontani
2nd Author's Affiliation Nagoya Institute of Technology(NIT)
3rd Author's Name Hiroshi Matsuzoe
3rd Author's Affiliation Nagoya Institute of Technology(NIT)
Date 2015-09-08
Paper # MI2015-54
Volume (vol) vol.115
Number (no) MI-218
Page pp.pp.31-34(MI),
#Pages 4
Date of Issue 2015-09-01 (MI)