Statistical pattern recognition techniques were applied to discriminate between healthy and flat foot children through ground reaction force (GRF) measurements, 28 young subjects (age 11.8±1.8) with symptomatic flexible flat foot, and 13 healthy subjects (age 10.1±1.4) took part on this preliminar study. The Karhunen-Loeve expansion with Fukunaga-Koontz normalization was applied to each GRF component to extract concise and discriminant features from the original data. The best performances of the classifying process were obtained using three principal components of both vertical and fore-aft GRF projections. Both linear and quadratic Bayes classifiers were designed with 9.2% and 4.9% total this misclassification errors for vertical, and 14.9% and 9.5% for fore-aft projections, respectively. Analysis of correlation between the patterns of each class and the first principal component demonstrates that the most representative factors in the functional diagnosis of flat foot are the less absorption in the fore-aft component and the reduction of the second vertical peak during push-off.

Flat foot classification using principal component analysis applied to ground reaction forces / Bertani, A.; Cappello, A.; Catani, F.; Benedetti, M. G.; Giannini, S.. - (1997), pp. 259-268. ((Intervento presentato al convegno Proceedings of the 1997 4th International Conference on Simulations in Biomedicine, BIOMED tenutosi a Acquasparta, Italy, nel 1997.

Flat foot classification using principal component analysis applied to ground reaction forces

Bertani A.;Catani F.;Benedetti M. G.;Giannini S.
1997-01-01

Abstract

Statistical pattern recognition techniques were applied to discriminate between healthy and flat foot children through ground reaction force (GRF) measurements, 28 young subjects (age 11.8±1.8) with symptomatic flexible flat foot, and 13 healthy subjects (age 10.1±1.4) took part on this preliminar study. The Karhunen-Loeve expansion with Fukunaga-Koontz normalization was applied to each GRF component to extract concise and discriminant features from the original data. The best performances of the classifying process were obtained using three principal components of both vertical and fore-aft GRF projections. Both linear and quadratic Bayes classifiers were designed with 9.2% and 4.9% total this misclassification errors for vertical, and 14.9% and 9.5% for fore-aft projections, respectively. Analysis of correlation between the patterns of each class and the first principal component demonstrates that the most representative factors in the functional diagnosis of flat foot are the less absorption in the fore-aft component and the reduction of the second vertical peak during push-off.
Proceedings of the 1997 4th International Conference on Simulations in Biomedicine, BIOMED
Acquasparta, Italy,
1997
259
268
Bertani, A.; Cappello, A.; Catani, F.; Benedetti, M. G.; Giannini, S.
Flat foot classification using principal component analysis applied to ground reaction forces / Bertani, A.; Cappello, A.; Catani, F.; Benedetti, M. G.; Giannini, S.. - (1997), pp. 259-268. ((Intervento presentato al convegno Proceedings of the 1997 4th International Conference on Simulations in Biomedicine, BIOMED tenutosi a Acquasparta, Italy, nel 1997.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

Licenza Creative Commons
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11380/1288015
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact