Certain investigation on automated cry pattern analysis for early detection of medical disorders using deep learning techniques

Abstract

newline Infants cry to express their feeling to its parents and caretakers. The cries express various discomforts and needs of baby to external world. However, interpreting the reasons behind the baby cries is significant in order to handle the baby needs properly. Based on the reason behind baby cry, these signals possess different patterns which can be used to retrieve useful features through automated classification algorithms. In contrast to normal infants cry classification, neonates or preterm babies in NICU and incubators cry often which implies the underlying physical and medical disorders. Early detection of cry patterns can provide useful insights regarding the health of baby which assists in timely diagnosis and treatment. A new paradigm for sophisticated automatic cry pattern identification is made possible by the combination of machine learning and signal processing methods. The proposed work suggests a two module framework to extract cry patterns of neonates in NICU using Deep Learning techniques.