In this work the Otsu method is applied to the Short-Term Energy measure (STE) histogram of the signal. The method provides the optimal threshold tu for the separation of a bimodal histogram minimizing the intra-class variance of the two resulting classes (e.g. “cry” and “noise”). However, in case of energy oscillation around the threshold level, several disjoint cry episodes are detected instead of a single one. This problem has been solved by iteratively applying the Otsu algorithm to the histogram of the “noise” class to detect a lower threshold tl, and applying a hysteresis thresholding where tu is the upper threshold and tl the lower threshold: when the STE of the signal overpasses tu a starting point is detected, when the STE of the signal falls down tl the ending point of the event is found. This procedure identifies two sets representing the starting and the ending points of the cry episodes.

Automatic extraction of cry episodes from newborn infant cry recordings / S. Orlandi ; C. Risaliti ; C. Manfredi ; L. Bocchi ; G. P. Donzelli. - STAMPA. - (2010), pp. 32-32. (Intervento presentato al convegno The 4th COST 2103 Advanced Voice Function Assessment Workshop tenutosi a York, UK nel 19-21 May 2010).

Automatic extraction of cry episodes from newborn infant cry recordings

ORLANDI, SILVIA;MANFREDI, CLAUDIA;BOCCHI, LEONARDO;DONZELLI, GIAN PAOLO
2010

Abstract

In this work the Otsu method is applied to the Short-Term Energy measure (STE) histogram of the signal. The method provides the optimal threshold tu for the separation of a bimodal histogram minimizing the intra-class variance of the two resulting classes (e.g. “cry” and “noise”). However, in case of energy oscillation around the threshold level, several disjoint cry episodes are detected instead of a single one. This problem has been solved by iteratively applying the Otsu algorithm to the histogram of the “noise” class to detect a lower threshold tl, and applying a hysteresis thresholding where tu is the upper threshold and tl the lower threshold: when the STE of the signal overpasses tu a starting point is detected, when the STE of the signal falls down tl the ending point of the event is found. This procedure identifies two sets representing the starting and the ending points of the cry episodes.
2010
The 4th COST 2103 Advanced Voice Function Assessment Workshop Clinical voice assessment in the future - new horizons
The 4th COST 2103 Advanced Voice Function Assessment Workshop
York, UK
S. Orlandi ; C. Risaliti ; C. Manfredi ; L. Bocchi ; G. P. Donzelli
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificatore per citare o creare un link a questa risorsa: https://hdl.handle.net/2158/675357
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact