Domestic Language Accent Detector Using MFCC and GMM
Roshni Rajan. K, Annie P Joseph · International Journal of Applied Engineering Research · 2020
In day today life, speech is commonly used for communication purpose.So that its recognition has an important role.There are already recognition systems for English and that has been used in different automation and assistance purposes.But in the case of local languages it seems to be difficult in recognition.In the case of Malayalam the usage of words, its pronunciation as well as the meaning vary with respect to the regions.So the same word has different meaning in different areas and hence it become complex.This paper aims to recognize different accents of Malayalam in accounts with the place of the user.Initially a set of training data is acquired from different areas and from the data Mel Frequency Cepstral Co-efficients are extracted as features.A Gaussian Mixture Model is made in accordance with the data which uses Energy Maximization Algorithm.It recognizes the spoken accents and can be used for various home automation and virtual assisting technologies.