Abstract:

Communication in different type in which crying is one of the forms. Infants can only express their feelings through crying only. Baby’s cry can be characterized according to its natural periodic tone and the change of voice. Detection of a baby cry in speech signals is a crucial step in applications like remote baby monitoring. This study of sound recognition involves feature extraction and classification by determining the sound pattern Signal processing of crying signal and make the pattern classification using machine learning algorithms. Mel frequency cepstral coefficient (MFCC) method of feature extraction uses cry signal and classify the tone patterns of the audio signal. In this we are trying to increase the accuracy of the classification using K-NN algorithm. The KNN classifier consistently yields better results compared to other classifiers.

Description:

Speech Recognition, Audio Processing, Baby Crying, Signal Patterns, Speech signal Processing, Feature Extraction, MFCC, K- Nearest Neighbor