Article Preview
TopBackground
A number of research work related to this line have been reported, whereby many of which are based on Artificial Neural Network (ANN) classification techniques. (Petroni, Malowany, Johnston, & Stevens,1995) for example, have used three different varieties of supervised ANN technique which include a simple feed-forward, a recurrent neural network (RNN) and a time-delay neural network (TDNN) in their infant cry classification system. In their study, they have attempted to recognize and classify three categories of cry, namely ‘pain’, ‘fear’ and ‘hunger’ and the results demonstrated that the highest classification rate was achieved by using feed-forward neural network. Another research work carried out by (Cano & Escobedo, 1999) used the Kohonen's self-organizing maps (SOM) which is basically a variety of unsupervised ANN technique to classify different infant cries. (Rosales-Pérez, Reyes-Garcia, Gonzalez, & Arch-Tirado, 2012) used Genetic Selection of a Fuzzy Model (GSFM) for classification of infant cry where GSFM selects a combination of feature selection methods, type of fuzzy processing, learning algorithm, and its associated parameters that best fit to the data and have obtained up to 99.42% in recognition accuracy. (Al-Azzawi, 2014) designed an automatic infant cry recognition system based on the fuzzy transform (F-transform) that classifies two different kinds of cries, which come from physiological status and medical disease, a supervised MLP scaled conjugate ANN was used and the classification accuracy obtained was 96%.