Speech Emotion Recognition Project

We have proposed a model to recognize emotion from speech which is in Hindi. The database for the speech emotion recognition system is the IIT (Kharagpur) Simulated Emotion Hindi Speech Corpus (IITKGP-SEHSC). Our approach uses the combination of mel frequency cepstral coefficients (MFCCs), chroma and mel spectrogram frequencies, to identify the underlying emotions. The proposed Multi-Layer Perceptron Classifier can classify the emotions - anger, disgust, fear, happy, sad, sarcastic and surprise. The model was able to analyze the tone and pitch of various audio clips of both, male and female voices and the final accuracy obtained with MLP Classifier was 81.52%.