Yasmina Kheddache1, Chakib Tadj2. 1. Department of Electrical Engineering, École de Technologie Supérieure, Montréal Quebec, Canada. Electronic address: yasmina.kheddache.1@etsmtl.net. 2. Department of Electrical Engineering, École de Technologie Supérieure, Montréal Quebec, Canada.
Abstract
OBJECTIVE: A new approach to the automatic quantification of the acoustic parameters of the cries of healthy newborns and newborns with pathologies is presented. The purpose of the present study was to examine the relationship between acoustic parameters and pathologies of interest to characterize healthy and pathologic cries of newborns. METHODS: Using MATLAB, this study included automatic estimation of F0, RF1, RF2, percentage and tuning duration, transition duration, RF2 slope, and RF1:RF2 ratio. The database used includes full- and pre-term newborns, healthy, and pathologic cries. It contains 3000 cry samples of 1-second duration from 65 newborn babies aged from 1 day to 1 month old. RESULTS: Statistical analysis results reveal that the distributions of these acoustic cry parameters depend on the pathology of newborn. In this work, we successfully identify the quantitative relationship between the acoustic cry characteristics we examined and the diseases we studied. CONCLUSIONS: Our deduction is that quantification of the variability of these parameters is useful for differentiating the cries of a healthy newborn from those of a newborn with a pathology, and that these data can be used for the early diagnosis of newborn diseases.
OBJECTIVE: A new approach to the automatic quantification of the acoustic parameters of the cries of healthy newborns and newborns with pathologies is presented. The purpose of the present study was to examine the relationship between acoustic parameters and pathologies of interest to characterize healthy and pathologic cries of newborns. METHODS: Using MATLAB, this study included automatic estimation of F0, RF1, RF2, percentage and tuning duration, transition duration, RF2 slope, and RF1:RF2 ratio. The database used includes full- and pre-term newborns, healthy, and pathologic cries. It contains 3000 cry samples of 1-second duration from 65 newborn babies aged from 1 day to 1 month old. RESULTS: Statistical analysis results reveal that the distributions of these acoustic cry parameters depend on the pathology of newborn. In this work, we successfully identify the quantitative relationship between the acoustic cry characteristics we examined and the diseases we studied. CONCLUSIONS: Our deduction is that quantification of the variability of these parameters is useful for differentiating the cries of a healthy newborn from those of a newborn with a pathology, and that these data can be used for the early diagnosis of newborn diseases.