Malayalam Kannur Dialect Speech Emotion Recognition using Machine Learning

Speech is the most natural form of communication among various modes such as verbal and non verbal communications. Speech include information about languages, Speaker, messages and emotions. Emotions make speech more expressive and effective. There are so many language based speech emotion recognition systems are available. And there are several Malayalam speech emotion recognition systems. But there is no speech emotion recognition system in kannur dialect, which is used in northern district of Kerala. So we implementing speech emotion recognition system in Malayalam of Kannur dialect. Currently there is no data set for our project. So we decide to create a dataset of kannur dialect. Here we use Mel- frequency spectral coefficient (MFCC), Chroma shift, Spectral centroid, Spectral bandwidth, rolloff, Zero crossing rate and mel spectrogram as the walls of the spectral features for extraction. The support vector machine (SVM) and Random forest algorithm are decided to use in our project for classification. The proposed system will provide a Novel speech emotion recognition suitable for malayalam kannur dialect