Grasp recognition with uncalibrated data gloves - A comparison of classification methods

Guido Heumer, Hani Ben Amor, Matthias Weber, Bernhard Jung

Research output: Chapter in Book/Report/Conference proceedingConference contribution

38 Citations (Scopus)

Abstract

This paper presents a comparison of various classification methods for the problem of recognizing grasp types involved in object manipulations performed with a data glove. Conventional wisdom holds that data gloves need calibration in order to obtain accurate results. However, calibration is a time-consuming process, inherently user-specific, and its results are often not perfect. In contrast, the present study aims at evaluating recognition methods that do not require prior calibration of the data glove, by using raw sensor readings as input features and mapping them directly to different categories of hand shapes. An experiment was carried out, where test persons wearing a data glove had to grasp physical objects of different shapes corresponding to the various grasp types of the Schlesinger taxonomy. The collected data was analyzed with 28 classifiers including different types of neural networks, decision trees, Bayes nets, and lazy learners. Each classifier was analyzed in six different settings, representing various application scenarios with differing generalization demands. The results of this work are twofold: (1) We show that a reasonably well to highly reliable recognition of grasp types can be achieved - depending on whether or not the glove user is among those training the classifier - even with uncalibrated data gloves. (2) We identify the best performing classification methods for recognition of various grasp types. To conclude, cumbersome calibration processes before productive usage of data gloves can be spared in many situations.

Original languageEnglish (US)
Title of host publicationProceedings - IEEE Virtual Reality
Pages19-26
Number of pages8
DOIs
StatePublished - 2007
Externally publishedYes
Event2007 IEEE Virtual Reality Conference, VR'07 - Charlotte, NC, United States
Duration: Mar 10 2007Mar 14 2007

Other

Other2007 IEEE Virtual Reality Conference, VR'07
CountryUnited States
CityCharlotte, NC
Period3/10/073/14/07

Fingerprint

Calibration
Classifiers
Taxonomies
Decision trees
Neural networks
Sensors
Experiments

Keywords

  • Calibration
  • Classification methods
  • Data glove
  • Grasp recognition

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Heumer, G., Ben Amor, H., Weber, M., & Jung, B. (2007). Grasp recognition with uncalibrated data gloves - A comparison of classification methods. In Proceedings - IEEE Virtual Reality (pp. 19-26). [4161001] https://doi.org/10.1109/VR.2007.352459

Grasp recognition with uncalibrated data gloves - A comparison of classification methods. / Heumer, Guido; Ben Amor, Hani; Weber, Matthias; Jung, Bernhard.

Proceedings - IEEE Virtual Reality. 2007. p. 19-26 4161001.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Heumer, G, Ben Amor, H, Weber, M & Jung, B 2007, Grasp recognition with uncalibrated data gloves - A comparison of classification methods. in Proceedings - IEEE Virtual Reality., 4161001, pp. 19-26, 2007 IEEE Virtual Reality Conference, VR'07, Charlotte, NC, United States, 3/10/07. https://doi.org/10.1109/VR.2007.352459
Heumer G, Ben Amor H, Weber M, Jung B. Grasp recognition with uncalibrated data gloves - A comparison of classification methods. In Proceedings - IEEE Virtual Reality. 2007. p. 19-26. 4161001 https://doi.org/10.1109/VR.2007.352459
Heumer, Guido ; Ben Amor, Hani ; Weber, Matthias ; Jung, Bernhard. / Grasp recognition with uncalibrated data gloves - A comparison of classification methods. Proceedings - IEEE Virtual Reality. 2007. pp. 19-26
@inproceedings{6dd1cb65ce684f55a3d79c871cbb3384,
title = "Grasp recognition with uncalibrated data gloves - A comparison of classification methods",
abstract = "This paper presents a comparison of various classification methods for the problem of recognizing grasp types involved in object manipulations performed with a data glove. Conventional wisdom holds that data gloves need calibration in order to obtain accurate results. However, calibration is a time-consuming process, inherently user-specific, and its results are often not perfect. In contrast, the present study aims at evaluating recognition methods that do not require prior calibration of the data glove, by using raw sensor readings as input features and mapping them directly to different categories of hand shapes. An experiment was carried out, where test persons wearing a data glove had to grasp physical objects of different shapes corresponding to the various grasp types of the Schlesinger taxonomy. The collected data was analyzed with 28 classifiers including different types of neural networks, decision trees, Bayes nets, and lazy learners. Each classifier was analyzed in six different settings, representing various application scenarios with differing generalization demands. The results of this work are twofold: (1) We show that a reasonably well to highly reliable recognition of grasp types can be achieved - depending on whether or not the glove user is among those training the classifier - even with uncalibrated data gloves. (2) We identify the best performing classification methods for recognition of various grasp types. To conclude, cumbersome calibration processes before productive usage of data gloves can be spared in many situations.",
keywords = "Calibration, Classification methods, Data glove, Grasp recognition",
author = "Guido Heumer and {Ben Amor}, Hani and Matthias Weber and Bernhard Jung",
year = "2007",
doi = "10.1109/VR.2007.352459",
language = "English (US)",
isbn = "1424409055",
pages = "19--26",
booktitle = "Proceedings - IEEE Virtual Reality",

}

TY - GEN

T1 - Grasp recognition with uncalibrated data gloves - A comparison of classification methods

AU - Heumer, Guido

AU - Ben Amor, Hani

AU - Weber, Matthias

AU - Jung, Bernhard

PY - 2007

Y1 - 2007

N2 - This paper presents a comparison of various classification methods for the problem of recognizing grasp types involved in object manipulations performed with a data glove. Conventional wisdom holds that data gloves need calibration in order to obtain accurate results. However, calibration is a time-consuming process, inherently user-specific, and its results are often not perfect. In contrast, the present study aims at evaluating recognition methods that do not require prior calibration of the data glove, by using raw sensor readings as input features and mapping them directly to different categories of hand shapes. An experiment was carried out, where test persons wearing a data glove had to grasp physical objects of different shapes corresponding to the various grasp types of the Schlesinger taxonomy. The collected data was analyzed with 28 classifiers including different types of neural networks, decision trees, Bayes nets, and lazy learners. Each classifier was analyzed in six different settings, representing various application scenarios with differing generalization demands. The results of this work are twofold: (1) We show that a reasonably well to highly reliable recognition of grasp types can be achieved - depending on whether or not the glove user is among those training the classifier - even with uncalibrated data gloves. (2) We identify the best performing classification methods for recognition of various grasp types. To conclude, cumbersome calibration processes before productive usage of data gloves can be spared in many situations.

AB - This paper presents a comparison of various classification methods for the problem of recognizing grasp types involved in object manipulations performed with a data glove. Conventional wisdom holds that data gloves need calibration in order to obtain accurate results. However, calibration is a time-consuming process, inherently user-specific, and its results are often not perfect. In contrast, the present study aims at evaluating recognition methods that do not require prior calibration of the data glove, by using raw sensor readings as input features and mapping them directly to different categories of hand shapes. An experiment was carried out, where test persons wearing a data glove had to grasp physical objects of different shapes corresponding to the various grasp types of the Schlesinger taxonomy. The collected data was analyzed with 28 classifiers including different types of neural networks, decision trees, Bayes nets, and lazy learners. Each classifier was analyzed in six different settings, representing various application scenarios with differing generalization demands. The results of this work are twofold: (1) We show that a reasonably well to highly reliable recognition of grasp types can be achieved - depending on whether or not the glove user is among those training the classifier - even with uncalibrated data gloves. (2) We identify the best performing classification methods for recognition of various grasp types. To conclude, cumbersome calibration processes before productive usage of data gloves can be spared in many situations.

KW - Calibration

KW - Classification methods

KW - Data glove

KW - Grasp recognition

UR - http://www.scopus.com/inward/record.url?scp=34547650835&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34547650835&partnerID=8YFLogxK

U2 - 10.1109/VR.2007.352459

DO - 10.1109/VR.2007.352459

M3 - Conference contribution

SN - 1424409055

SN - 9781424409051

SP - 19

EP - 26

BT - Proceedings - IEEE Virtual Reality

ER -