Candidates for synergies: Linear Discriminants versus principal components

Ramana Vinjamuri, Vrajeshri Patel, Michael Powell, Zhi Hong Mao, Nathan E Crone

Research output: Contribution to journalArticle

Abstract

Movement primitives or synergies have been extracted from human hand movements using several matrix factorization, dimensionality reduction, and classification methods. Principal component analysis (PCA) is widely used to obtain the first few significant eigenvectors of covariance that explain most of the variance of the data. Linear discriminant analysis (LDA) is also used as a supervised learning method to classify the hand postures corresponding to the objects grasped. Synergies obtained using PCA are principal component vectors aligned with dominant variances. On the other hand, synergies obtained using LDA are linear discriminant vectors that separate the groups of variances. In this paper, time varying kinematic synergies in the human hand grasping movements were extracted using these two diametrically opposite methods and were evaluated in reconstructing natural and American sign language (ASL) postural movements. We used an unsupervised LDA (ULDA) to extract linear discriminants. The results suggest that PCA outperformed LDA. The uniqueness, advantages, and disadvantages of each of these methods in representing high-dimensional hand movements in reduced dimensions were discussed.

Original languageEnglish (US)
Article number373957
JournalComputational Intelligence and Neuroscience
Volume2014
DOIs
StatePublished - 2014

Fingerprint

Synergy
Discriminant Analysis
Discriminant analysis
Principal Components
Discriminant
Hand
Principal Component Analysis
Principal component analysis
Sign Language
Supervised learning
Posture
Factorization
Biomechanical Phenomena
Eigenvalues and eigenfunctions
Kinematics
Grasping
Matrix Factorization
Learning
Dimensionality Reduction
Supervised Learning

ASJC Scopus subject areas

  • Computer Science(all)
  • Mathematics(all)
  • Neuroscience(all)
  • Medicine(all)

Cite this

Candidates for synergies : Linear Discriminants versus principal components. / Vinjamuri, Ramana; Patel, Vrajeshri; Powell, Michael; Mao, Zhi Hong; Crone, Nathan E.

In: Computational Intelligence and Neuroscience, Vol. 2014, 373957, 2014.

Research output: Contribution to journalArticle

Vinjamuri, Ramana ; Patel, Vrajeshri ; Powell, Michael ; Mao, Zhi Hong ; Crone, Nathan E. / Candidates for synergies : Linear Discriminants versus principal components. In: Computational Intelligence and Neuroscience. 2014 ; Vol. 2014.
@article{25d3c86062cb41d399a7ef13cca7ed54,
title = "Candidates for synergies: Linear Discriminants versus principal components",
abstract = "Movement primitives or synergies have been extracted from human hand movements using several matrix factorization, dimensionality reduction, and classification methods. Principal component analysis (PCA) is widely used to obtain the first few significant eigenvectors of covariance that explain most of the variance of the data. Linear discriminant analysis (LDA) is also used as a supervised learning method to classify the hand postures corresponding to the objects grasped. Synergies obtained using PCA are principal component vectors aligned with dominant variances. On the other hand, synergies obtained using LDA are linear discriminant vectors that separate the groups of variances. In this paper, time varying kinematic synergies in the human hand grasping movements were extracted using these two diametrically opposite methods and were evaluated in reconstructing natural and American sign language (ASL) postural movements. We used an unsupervised LDA (ULDA) to extract linear discriminants. The results suggest that PCA outperformed LDA. The uniqueness, advantages, and disadvantages of each of these methods in representing high-dimensional hand movements in reduced dimensions were discussed.",
author = "Ramana Vinjamuri and Vrajeshri Patel and Michael Powell and Mao, {Zhi Hong} and Crone, {Nathan E}",
year = "2014",
doi = "10.1155/2014/373957",
language = "English (US)",
volume = "2014",
journal = "Computational Intelligence and Neuroscience",
issn = "1687-5265",
publisher = "Hindawi Publishing Corporation",

}

TY - JOUR

T1 - Candidates for synergies

T2 - Linear Discriminants versus principal components

AU - Vinjamuri, Ramana

AU - Patel, Vrajeshri

AU - Powell, Michael

AU - Mao, Zhi Hong

AU - Crone, Nathan E

PY - 2014

Y1 - 2014

N2 - Movement primitives or synergies have been extracted from human hand movements using several matrix factorization, dimensionality reduction, and classification methods. Principal component analysis (PCA) is widely used to obtain the first few significant eigenvectors of covariance that explain most of the variance of the data. Linear discriminant analysis (LDA) is also used as a supervised learning method to classify the hand postures corresponding to the objects grasped. Synergies obtained using PCA are principal component vectors aligned with dominant variances. On the other hand, synergies obtained using LDA are linear discriminant vectors that separate the groups of variances. In this paper, time varying kinematic synergies in the human hand grasping movements were extracted using these two diametrically opposite methods and were evaluated in reconstructing natural and American sign language (ASL) postural movements. We used an unsupervised LDA (ULDA) to extract linear discriminants. The results suggest that PCA outperformed LDA. The uniqueness, advantages, and disadvantages of each of these methods in representing high-dimensional hand movements in reduced dimensions were discussed.

AB - Movement primitives or synergies have been extracted from human hand movements using several matrix factorization, dimensionality reduction, and classification methods. Principal component analysis (PCA) is widely used to obtain the first few significant eigenvectors of covariance that explain most of the variance of the data. Linear discriminant analysis (LDA) is also used as a supervised learning method to classify the hand postures corresponding to the objects grasped. Synergies obtained using PCA are principal component vectors aligned with dominant variances. On the other hand, synergies obtained using LDA are linear discriminant vectors that separate the groups of variances. In this paper, time varying kinematic synergies in the human hand grasping movements were extracted using these two diametrically opposite methods and were evaluated in reconstructing natural and American sign language (ASL) postural movements. We used an unsupervised LDA (ULDA) to extract linear discriminants. The results suggest that PCA outperformed LDA. The uniqueness, advantages, and disadvantages of each of these methods in representing high-dimensional hand movements in reduced dimensions were discussed.

UR - http://www.scopus.com/inward/record.url?scp=84927170534&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84927170534&partnerID=8YFLogxK

U2 - 10.1155/2014/373957

DO - 10.1155/2014/373957

M3 - Article

C2 - 25143763

AN - SCOPUS:84927170534

VL - 2014

JO - Computational Intelligence and Neuroscience

JF - Computational Intelligence and Neuroscience

SN - 1687-5265

M1 - 373957

ER -