New Born Baby Cry Analysis and Classification

Abstract

This study proposes a baby cry detection system using Mel Frequency Cepstral Coefficients (MFCC) and K-Nearest Neighbor (KNN) algorithm. The review of a wide variety of literature focuses primarily on data gathering, processing of signals in cross domain, and classification using machine learning algorithms. When using various apps to monitor a baby’s condition, automatic voice detection of a baby’s cry is crucial. This hypothesized concept involves the detection of a baby’s scream. To distinguish infant sounds of crying in a range of settings which are residential under challenging circumstances, this system uses a machine learning technique. The proposed method extracts MFCC features from the audio signals of the baby cries and applies KNN to classify the cries based on their features. The results are based on a classification-related investigation of the performance of cry-detection using KNN. A baby’s cry may be automatically detected in a few circumstances including different environmental noises. The suggested technique demonstrated high accuracy in identifying cry types when tested using a freely accessible dataset.