Infant Cries Recognition using Neural Network / Muhammad Fariduddin Mohd Yasin

• Main Author: Mohd Yasin, Muhammad Fariduddin
• Format: Thesis
• Language: English
• Published: Universiti Teknologi MARA, Shah Alam 2015
• Subjects: Electronic Computers. Computer Science; Android; Neural networks (Computer science)
• Online Access: https://ir.uitm.edu.my/id/eprint/49255/1/49255.pdf; https://ir.uitm.edu.my/id/eprint/49255/

In this work, a study to recognize and classify the infant's cries was presented as most of parents or people cannot understand why their baby is crying. The information and knowledge is acquitted from several documents and researches done by other researchers. The method used in order to recognize and classify the cries is Neural Network. The classification is done using the software called Matlab, where the crying sounds during the "pre-cry stage" that is before the babies start crying hysterically are taken as the input. The cries were recorded using the function in the Matlab and Neural Network is applied. The cries are classified based on research made by baby language's researchers, Priscilla Dunstan. This study gets the data from several videos of Priscilla's researches and trained it. The trained data then is tested with new recorded cries. As a result, this study successfully recognizes several infant cries based on its types and it is hoped that this type of study is conducted on a bigger scale.