Infant Cry Detection and Pain Scale Assessment: A Pilot Study

2.1. Infant Cry

The infant cry signal used for the study is taken from Chillanto, Mexican data base which is used for the study of characteristics of the cry with conditions such as normal, pain, hunger, deaf and Asphyxia. The full recording and the segmented of each recording are used as the preliminary study infant cry recordings. The data base gives us the details about all the mentioned cry and these cries can be used reference for the recorded cry in this project. It is similar to Automatic Speech Recognition system process which comprises two main stages. They are signal processing and pattern classification. The main aim of this system is to discriminate the different types of cries and identify the pathology cry. In signal processing step raw cry signal is preprocessed such as cleaning, filtering and normalization. Then smoothed signal is analyzed and by suitable feature extraction techniques the salient features are derived.