Dimensionality reduction of electropalatographic data using latent variable models

Carreira-Perpiñán, Miguel Á.; Renals, Steve

doi:10.1016/s0167-6393(98)00059-4

Cited by 14 publications

(11 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Contrarily, PFA determines the least number of factors that can account for the common variance (correlation). For more details and illustrative examples on PCA and PFA, please refer to [7,11,12].…”

Section: A Model Constructionmentioning

confidence: 99%

Interpretability of anatomical variability analysis of abdominal organs via clusterization of decomposition modes

Ballester¹,

Kozic

et al. 2008

2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society

View full text Add to dashboard Cite

Extensive recent work has taken place on the construction of probabilistic atlases of anatomical organs, especially the brain, and their application in medical image analysis. These techniques are leading the way into similar studies of other organs and more comprehensively of groups of organs. In this paper we report results on the analysis of anatomical variability obtained from probabilistic atlases of abdominal organs. Two factor analysis techniques, namely principal component analysis (PCA) and principal factor analysis (PFA), were used to decompose and study shape variability within the abdomen. To assess and ease the interpretability of the resulting deformation modes, a clustering technique of the deformation vectors is proposed. The analysis of deformation fields obtained using these two factor analysis techniques showed strong correlation with anatomical landmarks and known mechanical deformations in the abdomen, allowing us to conclude that PFA is a complementary decomposition technique that offers easy-to-interpret additional information to PCA in a clinical setting. The analysis of organ anatomical variability will represent a potentially important research tool for abdominal diagnosis and modeling.

show abstract

Section: A Model Constructionmentioning

confidence: 99%

Interpretability of anatomical variability analysis of abdominal organs via clusterization of decomposition modes

Ballester¹,

Kozic

et al. 2008

2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society

View full text Add to dashboard Cite

show abstract

“…Generative models try to model the density function that is assumed to have generated the data, under constraints that restricts the set of possible models to those with low intrinsic dimensionality. The following description is mainly based on Carreira-Perpiñán's paper [4]. PFA represents an observed D-dimensional continuous variable t, as a linear function f of an L-dimensional (L<D) continuous latent variable x and an independent Gaussian noise process:…”

Section: Principal Factor Analysismentioning

confidence: 99%

“…PCA belongs to a family of methods for multivariate analysis commonly known as Factor Analysis (FA). Reviews and comparative studies of FA techniques can be found elsewhere [3] [4]. Such techniques can be classified into linear and non-linear, reflecting whether the shape variation can be expressed as a linear combination of basic deformation primitives.…”

Section: Introductionmentioning

confidence: 99%

Statistical Shape Analysis via Principal Factor Analysis

Aguirre

Linguraru

Marias

et al. 2007

2007 4th IEEE International Symposium on Biomedical Imaging: From Nano to Macro

View full text Add to dashboard Cite

Statistical shape analysis techniques commonly employed in the medical imaging community, such as Active Shape Models or Active Appearance Models, rely on Principal Component Analysis (PCA) to decompose shape variability into a reduced set of interpretable components. In this paper we propose Principal Factor Analysis (PFA) as an alternative and complementary tool to PCA providing a decomposition into modes of variation that can be more easily interpretable, while still being a linear efficient technique that performs dimensionality reduction (as opposed to Independent Component Analysis, ICA). The key difference between PFA and PCA is that PFA models covariance between variables, rather than the total variance in the data. The added value of PFA is illustrated on 2D landmark data of corpora callosa outlines. Then, a study of the 3D shape variability of the human left femur is performed. Finally, we report results on vector-valued 3D deformation fields resulting from non-rigid registration of ventricles in MRI of the brain.

show abstract

“…Reviews and comparative studies of FA techniques can be found elsewhere. 3,7,16 Such techniques can be classified into linear and non-linear, reflecting whether the shape variation can be expressed as a linear combination of basic deformation primitives. We contend that a factorial decomposition of shape variability, if it is to be easily interpretable, must follow a linear model, where each mode of variation has a scalar weight.…”

Section: Introductionmentioning

confidence: 99%

On the adequacy of principal factor analysis for the study of shape variability

et al. 2005

View full text Add to dashboard Cite

The analysis of shape variability of anatomical structures is of key importance in a number of clinical disciplines, as abnormality in shape can be related to certain diseases. Statistical shape analysis techniques commonly employed in the medical imaging community, such as Active Shape Models or Active Appearance Models rely on Principal Component Analysis (PCA) to decompose shape variability into a reduced set of interpretable components. In this paper we propose Principal Factor Analysis (PFA) as an alternative to PCA and argue that PFA is a better suited technique for medical imaging applications. PFA provides a decomposition into modes of variation that are more easily interpretable, while still being a linear, efficient technique that performs dimensionality reduction (as opposed to Independent Component Analysis, ICA). Both PCA and PFA are described. Examples are provided for 2D landmark data of corpora callosa outlines, as well as vector-valued 3D deformation fields resulting from non-rigid registration of ventricles in MRI. The results show that PFA is a more descriptive tool for shape analysis, at a small cost in size (as in theory more components may be necessary to explain a given percentage of total variance in the data). In conclusion, we argue that it is important to study the potential of factor analysis techniques other than PCA for the application of shape analysis, and defend PFA as a good alternative.

show abstract

Dimensionality reduction of electropalatographic data using latent variable models

Cited by 14 publications

References 30 publications

Interpretability of anatomical variability analysis of abdominal organs via clusterization of decomposition modes

Interpretability of anatomical variability analysis of abdominal organs via clusterization of decomposition modes

Statistical Shape Analysis via Principal Factor Analysis

On the adequacy of principal factor analysis for the study of shape variability

Contact Info

Product

Resources

About