Morphological appearance manifolds for group-wise morphometric analysis

Nai Xiang Lian, Christos Davatzikos

Research output: Contribution to journalArticle

Abstract

Computational anatomy quantifies anatomical shape based on diffeomorphic transformations of a template. However, different templates warping algorithms, regularization parameters, or templates, lead to different representations of the same exact anatomy, raising a uniqueness issue: variations of these parameters are confounding factors as they give rise to non-unique representations. Recently, it has been shown that learning the equivalence class derived from the multitude of representations of a given anatomy can lead to improved and more stable morphological descriptors. Herein, we follow that approach, by approximating this equivalence class of morphological descriptors by a (nonlinear) morphological appearance manifold fitting to the data via a locally linear model. Our approach parallels work in the computer vision field, in which variations lighting, pose and other parameters lead to image appearance manifolds representing the exact same figure in different ways. The proposed framework is then used for group-wise registration and statistical analysis of biomedical images, by employing a minimum variance criterion to perform manifold-constrained optimization, i.e. to traverse each individual's morphological appearance manifold until group variance is minimal. The hypothesis is that this process is likely to reduce aforementioned confounding effects and potentially lead to morphological representations reflecting purely biological variations, instead of variations introduced by modeling assumptions and parameter settings.

Original languageEnglish (US)
Pages (from-to)814-829
Number of pages16
JournalMedical Image Analysis
Volume15
Issue number6
DOIs
StatePublished - Dec 2011
Externally publishedYes

Fingerprint

Equivalence classes
Anatomy
Constrained optimization
Computer vision
Statistical methods
Lighting
Linear Models
Learning

Keywords

  • Computational anatomy
  • Morphological appearance manifolds
  • Nonlinear representation
  • Spatial-varying optimization

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition
  • Radiology Nuclear Medicine and imaging
  • Health Informatics
  • Radiological and Ultrasound Technology

Cite this

Morphological appearance manifolds for group-wise morphometric analysis. / Lian, Nai Xiang; Davatzikos, Christos.

In: Medical Image Analysis, Vol. 15, No. 6, 12.2011, p. 814-829.

Research output: Contribution to journalArticle

Lian, Nai Xiang ; Davatzikos, Christos. / Morphological appearance manifolds for group-wise morphometric analysis. In: Medical Image Analysis. 2011 ; Vol. 15, No. 6. pp. 814-829.
@article{7156d4c8471b4bef8a3ac22851d9c82a,
title = "Morphological appearance manifolds for group-wise morphometric analysis",
abstract = "Computational anatomy quantifies anatomical shape based on diffeomorphic transformations of a template. However, different templates warping algorithms, regularization parameters, or templates, lead to different representations of the same exact anatomy, raising a uniqueness issue: variations of these parameters are confounding factors as they give rise to non-unique representations. Recently, it has been shown that learning the equivalence class derived from the multitude of representations of a given anatomy can lead to improved and more stable morphological descriptors. Herein, we follow that approach, by approximating this equivalence class of morphological descriptors by a (nonlinear) morphological appearance manifold fitting to the data via a locally linear model. Our approach parallels work in the computer vision field, in which variations lighting, pose and other parameters lead to image appearance manifolds representing the exact same figure in different ways. The proposed framework is then used for group-wise registration and statistical analysis of biomedical images, by employing a minimum variance criterion to perform manifold-constrained optimization, i.e. to traverse each individual's morphological appearance manifold until group variance is minimal. The hypothesis is that this process is likely to reduce aforementioned confounding effects and potentially lead to morphological representations reflecting purely biological variations, instead of variations introduced by modeling assumptions and parameter settings.",
keywords = "Computational anatomy, Morphological appearance manifolds, Nonlinear representation, Spatial-varying optimization",
author = "Lian, {Nai Xiang} and Christos Davatzikos",
year = "2011",
month = "12",
doi = "10.1016/j.media.2011.06.003",
language = "English (US)",
volume = "15",
pages = "814--829",
journal = "Medical Image Analysis",
issn = "1361-8415",
publisher = "Elsevier",
number = "6",

}

TY - JOUR

T1 - Morphological appearance manifolds for group-wise morphometric analysis

AU - Lian, Nai Xiang

AU - Davatzikos, Christos

PY - 2011/12

Y1 - 2011/12

N2 - Computational anatomy quantifies anatomical shape based on diffeomorphic transformations of a template. However, different templates warping algorithms, regularization parameters, or templates, lead to different representations of the same exact anatomy, raising a uniqueness issue: variations of these parameters are confounding factors as they give rise to non-unique representations. Recently, it has been shown that learning the equivalence class derived from the multitude of representations of a given anatomy can lead to improved and more stable morphological descriptors. Herein, we follow that approach, by approximating this equivalence class of morphological descriptors by a (nonlinear) morphological appearance manifold fitting to the data via a locally linear model. Our approach parallels work in the computer vision field, in which variations lighting, pose and other parameters lead to image appearance manifolds representing the exact same figure in different ways. The proposed framework is then used for group-wise registration and statistical analysis of biomedical images, by employing a minimum variance criterion to perform manifold-constrained optimization, i.e. to traverse each individual's morphological appearance manifold until group variance is minimal. The hypothesis is that this process is likely to reduce aforementioned confounding effects and potentially lead to morphological representations reflecting purely biological variations, instead of variations introduced by modeling assumptions and parameter settings.

AB - Computational anatomy quantifies anatomical shape based on diffeomorphic transformations of a template. However, different templates warping algorithms, regularization parameters, or templates, lead to different representations of the same exact anatomy, raising a uniqueness issue: variations of these parameters are confounding factors as they give rise to non-unique representations. Recently, it has been shown that learning the equivalence class derived from the multitude of representations of a given anatomy can lead to improved and more stable morphological descriptors. Herein, we follow that approach, by approximating this equivalence class of morphological descriptors by a (nonlinear) morphological appearance manifold fitting to the data via a locally linear model. Our approach parallels work in the computer vision field, in which variations lighting, pose and other parameters lead to image appearance manifolds representing the exact same figure in different ways. The proposed framework is then used for group-wise registration and statistical analysis of biomedical images, by employing a minimum variance criterion to perform manifold-constrained optimization, i.e. to traverse each individual's morphological appearance manifold until group variance is minimal. The hypothesis is that this process is likely to reduce aforementioned confounding effects and potentially lead to morphological representations reflecting purely biological variations, instead of variations introduced by modeling assumptions and parameter settings.

KW - Computational anatomy

KW - Morphological appearance manifolds

KW - Nonlinear representation

KW - Spatial-varying optimization

UR - http://www.scopus.com/inward/record.url?scp=80053969983&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80053969983&partnerID=8YFLogxK

U2 - 10.1016/j.media.2011.06.003

DO - 10.1016/j.media.2011.06.003

M3 - Article

VL - 15

SP - 814

EP - 829

JO - Medical Image Analysis

JF - Medical Image Analysis

SN - 1361-8415

IS - 6

ER -