A Survey on Detection of Reasons Behind Infant Cry Using Speech Processing

Infant cry analysis is necessary to detect the reasons behind the cry. Infants can communicate through their cry only. They express their physical and emotional needs through their cry. The crying of an infant can be stopped if the reason behind their cry is known. Therefore, identifying the reason behind infant cry is very essential for medical follow up. In this paper, a literature survey is done to study, analyze, compare the existing approaches and their limitations. This survey helped us to propose a new approach for detecting the reasons behind Infant's cry. The proposed model will use input as the cry signal and from these signal patterns the unique features are extracted using MFCCs, LPCCs and Pitch, etc. These features will help discriminating the patterns from one another. These extracted features will be used to train Neural Network Multilayer classifier. This classifier will be used to identify multiple classes of reasons behind infant cry such as hunger, pain, sleep, discomfort, etc. Efforts will be made for achieving good accuracy results through this approach.