Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand Kinematics
The purpose of this study was to find a parsimonious representation of hand kinematics data that could facilitate prosthetic hand control. Principal Component Analysis (PCA) and a non-linear Autoencoder Network (nAEN) were compared in their effectiveness at capturing the essential characteristics of...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2020-05-01
|
| Series: | Frontiers in Bioengineering and Biotechnology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/article/10.3389/fbioe.2020.00429/full |
| _version_ | 1811223963298168832 |
|---|---|
| author | Alexandra A. Portnova-Fahreeva Alexandra A. Portnova-Fahreeva Fabio Rizzoglio Fabio Rizzoglio Fabio Rizzoglio Ilana Nisky Maura Casadio Maura Casadio Ferdinando A. Mussa-Ivaldi Ferdinando A. Mussa-Ivaldi Ferdinando A. Mussa-Ivaldi Eric Rombokas Eric Rombokas |
| author_facet | Alexandra A. Portnova-Fahreeva Alexandra A. Portnova-Fahreeva Fabio Rizzoglio Fabio Rizzoglio Fabio Rizzoglio Ilana Nisky Maura Casadio Maura Casadio Ferdinando A. Mussa-Ivaldi Ferdinando A. Mussa-Ivaldi Ferdinando A. Mussa-Ivaldi Eric Rombokas Eric Rombokas |
| author_sort | Alexandra A. Portnova-Fahreeva |
| collection | DOAJ |
| description | The purpose of this study was to find a parsimonious representation of hand kinematics data that could facilitate prosthetic hand control. Principal Component Analysis (PCA) and a non-linear Autoencoder Network (nAEN) were compared in their effectiveness at capturing the essential characteristics of a wide spectrum of hand gestures and actions. Performance of the two methods was compared on (a) the ability to accurately reconstruct hand kinematic data from a latent manifold of reduced dimension, (b) variance distribution across latent dimensions, and (c) the separability of hand movements in compressed and reconstructed representations derived using a linear classifier. The nAEN exhibited higher performance than PCA in its ability to more accurately reconstruct hand kinematic data from a latent manifold of reduced dimension. Whereas, for two dimensions in the latent manifold, PCA was able to account for 78% of input data variance, nAEN accounted for 94%. In addition, the nAEN latent manifold was spanned by coordinates with more uniform share of signal variance compared to PCA. Lastly, the nAEN was able to produce a manifold of more separable movements than PCA, as different tasks, when reconstructed, were more distinguishable by a linear classifier, SoftMax regression. It is concluded that non-linear dimensionality reduction may offer a more effective platform than linear methods to control prosthetic hands. |
| first_indexed | 2024-04-12T08:42:16Z |
| format | Article |
| id | doaj.art-e1baf688028a4c0aaf4610a924fbde02 |
| institution | Directory Open Access Journal |
| issn | 2296-4185 |
| language | English |
| last_indexed | 2024-04-12T08:42:16Z |
| publishDate | 2020-05-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Bioengineering and Biotechnology |
| spelling | doaj.art-e1baf688028a4c0aaf4610a924fbde022022-12-22T03:39:52ZengFrontiers Media S.A.Frontiers in Bioengineering and Biotechnology2296-41852020-05-01810.3389/fbioe.2020.00429532101Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand KinematicsAlexandra A. Portnova-Fahreeva0Alexandra A. Portnova-Fahreeva1Fabio Rizzoglio2Fabio Rizzoglio3Fabio Rizzoglio4Ilana Nisky5Maura Casadio6Maura Casadio7Ferdinando A. Mussa-Ivaldi8Ferdinando A. Mussa-Ivaldi9Ferdinando A. Mussa-Ivaldi10Eric Rombokas11Eric Rombokas12Department of Mechanical Engineering, Northwestern University, Evanston, IL, United StatesShirley Ryan Ability Lab, Chicago, IL, United StatesShirley Ryan Ability Lab, Chicago, IL, United StatesDepartment of Physiology, Feinberg School of Medicine, Northwestern University, Chicago, IL, United StatesDepartment of Informatics, Bioengineering, Robotics and Systems Engineering, University of Genoa, Genoa, ItalyDepartment of Biomedical Engineering, Ben-Gurion University of the Negev, Be'er Sheva, IsraelDepartment of Physiology, Feinberg School of Medicine, Northwestern University, Chicago, IL, United StatesDepartment of Informatics, Bioengineering, Robotics and Systems Engineering, University of Genoa, Genoa, ItalyDepartment of Mechanical Engineering, Northwestern University, Evanston, IL, United StatesShirley Ryan Ability Lab, Chicago, IL, United StatesDepartment of Physiology, Feinberg School of Medicine, Northwestern University, Chicago, IL, United StatesDepartment of Mechanical Engineering, University of Washington, Seattle, WA, United StatesDepartment of Electrical Engineering, University of Washington, Seattle, WA, United StatesThe purpose of this study was to find a parsimonious representation of hand kinematics data that could facilitate prosthetic hand control. Principal Component Analysis (PCA) and a non-linear Autoencoder Network (nAEN) were compared in their effectiveness at capturing the essential characteristics of a wide spectrum of hand gestures and actions. Performance of the two methods was compared on (a) the ability to accurately reconstruct hand kinematic data from a latent manifold of reduced dimension, (b) variance distribution across latent dimensions, and (c) the separability of hand movements in compressed and reconstructed representations derived using a linear classifier. The nAEN exhibited higher performance than PCA in its ability to more accurately reconstruct hand kinematic data from a latent manifold of reduced dimension. Whereas, for two dimensions in the latent manifold, PCA was able to account for 78% of input data variance, nAEN accounted for 94%. In addition, the nAEN latent manifold was spanned by coordinates with more uniform share of signal variance compared to PCA. Lastly, the nAEN was able to produce a manifold of more separable movements than PCA, as different tasks, when reconstructed, were more distinguishable by a linear classifier, SoftMax regression. It is concluded that non-linear dimensionality reduction may offer a more effective platform than linear methods to control prosthetic hands.https://www.frontiersin.org/article/10.3389/fbioe.2020.00429/fullkinematicsneural networksprincipal component analysisdimensionality reductionunsupervised learningprosthetics |
| spellingShingle | Alexandra A. Portnova-Fahreeva Alexandra A. Portnova-Fahreeva Fabio Rizzoglio Fabio Rizzoglio Fabio Rizzoglio Ilana Nisky Maura Casadio Maura Casadio Ferdinando A. Mussa-Ivaldi Ferdinando A. Mussa-Ivaldi Ferdinando A. Mussa-Ivaldi Eric Rombokas Eric Rombokas Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand Kinematics Frontiers in Bioengineering and Biotechnology kinematics neural networks principal component analysis dimensionality reduction unsupervised learning prosthetics |
| title | Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand Kinematics |
| title_full | Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand Kinematics |
| title_fullStr | Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand Kinematics |
| title_full_unstemmed | Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand Kinematics |
| title_short | Linear and Non-linear Dimensionality-Reduction Techniques on Full Hand Kinematics |
| title_sort | linear and non linear dimensionality reduction techniques on full hand kinematics |
| topic | kinematics neural networks principal component analysis dimensionality reduction unsupervised learning prosthetics |
| url | https://www.frontiersin.org/article/10.3389/fbioe.2020.00429/full |
| work_keys_str_mv | AT alexandraaportnovafahreeva linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT alexandraaportnovafahreeva linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT fabiorizzoglio linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT fabiorizzoglio linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT fabiorizzoglio linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT ilananisky linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT mauracasadio linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT mauracasadio linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT ferdinandoamussaivaldi linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT ferdinandoamussaivaldi linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT ferdinandoamussaivaldi linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT ericrombokas linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics AT ericrombokas linearandnonlineardimensionalityreductiontechniquesonfullhandkinematics |