Emotion detection from Audio

Emotions are part of part of daily life for human beings, it serves major role on interactions. While making an interactions, finding emotions are essential. Speech Emotion Recognition (SER) is major area of research as it is challenging as it may change with time and can be recognized based on studying feature very closely. Speech emotions has variations in speech signals and pitch values for every different emotions. Supervised dataset is considered for learning speech emotions with different class values represents different emotions. Feature extraction plays major role in classification of emotions from dataset. With the growing computer vision applications, there is demand for SER researches to get the emotions accurately. The proposed work used Convolutional Neural Network (CNN) algorithm for learning and classification of supervised dataset set of audio emotions are considered as multi class dataset with emotions namely happiness, surprise, anger, neutral state, sadness, etc. Experimental results given proven results of emotion recognition and the live emotion recognition is achieved in the proposed work.