AUTOMATIC CLASSIFICATION OF THE CRY OF INFANTS WITH CLEFT PALATE

It is well known that the cry sound produced by an infant conveys information concerning the infant’s health or needs. In this paper, an Hidden Markov Model (HMM) based system for classification and monitoring of infants cry is presented. Classification tests among seven infants with cleft palate, with/without a palatal plate were performed in three forms: (1) age and subject dependent (2) age dependent – subject independent (3) age independent – subject dependent. The results were: 90%, 58% and 83% correct classification for the 1-3 case, respectively.

[1]  R H Colton,et al.  The cry characteristics of an infant who died of the sudden infant death syndrome. , 1981, The Journal of speech and hearing disorders.

[2]  K. Michelsson,et al.  Sound Spectrographic Cry Analysis of Infants with Hydrocephalus , 1984, Acta paediatrica Scandinavica.

[3]  K. Wermke,et al.  Spectral Analysis of Prespeech Sounds (Spontaneous Cries) in Infants with Unilateral Cleft Lip and Palate (UCLP): A Pilot Study , 2002 .

[4]  Arnon D. Cohen,et al.  Hidden Markov models in biomedical signal processing , 1998, Proceedings of the 20th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. Vol.20 Biomedical Engineering Towards the Year 2000 and Beyond (Cat. No.98CH36286).