Multiresolution for the detection of xiphoidal sounds in noisy medical audio signals

Abstract : In this paper a method we have developed in order to detect particular events (xiphoidal sounds) in swallowing audio signals is presented. The difficulty of the problem comes from the noisy nature of our medical audio signals, due to biological and environmental reasons. Besides, xiphoidal sounds have a highly nonstationary behaviour that makes conventional analysis methods such as Fourier transform or parametric model based methods nonsuitable. Thus, the detection method we have developed is based on a wavelet multiresolution analysis of the signal which is a powerful concept particularly adapted to signals dominated by transient behaviour or discontinuities. The decomposition levels of signals that comprise significant information are automatically selected and a multicriteria decision function is used to detect xiphoidal sounds among selected peaks.
Type de document :
Communication dans un congrès
4th EURASIP Conference focused on Video/Image Processing and Multimedia Communication, Jul 2003, Zagreb, Croatia. IEEE, pp.619-624, 2003, 〈10.1109/VIPMC.2003.1220532〉
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01610978
Contributeur : Équipe Gestionnaire Des Publications Si Liris <>
Soumis le : jeudi 5 octobre 2017 - 11:58:46
Dernière modification le : mercredi 21 mars 2018 - 10:54:07

Identifiants

Collections

Citation

Emmanuel Dellandréa, Pascal Makris, Michèle Boiron, Nicole Vincent. Multiresolution for the detection of xiphoidal sounds in noisy medical audio signals. 4th EURASIP Conference focused on Video/Image Processing and Multimedia Communication, Jul 2003, Zagreb, Croatia. IEEE, pp.619-624, 2003, 〈10.1109/VIPMC.2003.1220532〉. 〈hal-01610978〉

Partager

Métriques

Consultations de la notice

51