An intelligent system for infant cry detection and information in real time

Abstract

Acoustic analysis of infant cry signal can indicate the physical-emotional state of a baby and the cry-cause. Knowing it remotely is a challenge for working parents. This paper proposes an `Intelligent Cry Detection Information System (ICDIS)', to automatically detect an infant's cry and alert the parents. Parents can verify whether baby is in-fact crying, by informing a care taker who can send a web/app-controlled smart robot to fetch live video-streaming. It also sends text/image message to the parent, with urgency-level. The proposed system consists of sensors, Signal Processing Unit (SPU) and Information Transmission Unit (ITU). The audio-visual signals are processed in SPU. Acoustic features extracted from cry signal are analyzed. If a cry is detected then ITU automatically sends SMS to the parents. Performance of the prototype ICDIS is evaluated using a `Infant Cry Signals Database'. The proposed system has potential applications in child care centers, schools and hospitals etc.