A low cost intelligent smart system for real time infant monitoring and cry detection

摘要

Acoustic analysis of cry signals helps understanding the physiological and emotional state of an infant, and the likely cause of cry. In our earlier studies, acoustic analysis was carried out to characterize the infant cry signal due to different cause of crying. In this paper, we propose an Intelligent Cry Monitoring and Cry Detection System (ICMCDS) that can help working parents who are constrained to leave their children with the care-takers, and remain worried about their well-being. They need to be notified when their child cries. Functional prototype of a smart ICMCDS is developed at low-cost, which can monitor a child in real-time. Signal processing methods auto-correlation and a recently proposed modified zero-frequency filtering are used for cry-detection. Features F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> , Strength of Excitation and Signal Energy are extracted from the cry signal. Parameters mean, standard deviation and normalized standard deviation in F <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> are derived. Acoustic analysis of the cry signal helps detecting a cry as well as the cry-cause. The system detects a cry and informs the parent automatically through sms. If the child-cry continues for some time, then the intelligent system sends a second message, raising urgency concern. Further the system also identifies the likely cause of crying by examining the features, and informs to the parents by a third sms. The concerned parent may talk to a human controller, who in-turn can give control commands to a robot, using a webpage or mobile phone. The robot, equipped with audio-visual sensors, can be sent near to the child to get the actual state, by sending back live video streaming of the child and surroundings. It also validates the alert message about child crying. Wi-Fi medium is used for communication between humans and the robot. Performance evaluation is carried out using an Infant Cry Signals Database (IIITS ICSD 2). The proposed ICMCDS has a range of applications in homes, hospitals, child-care centers and schools.