Speech Emotion Features Selection using Multilayer Perceptron

Emotion recognition is the concept of identifying human feelings. People vary widely in their accuracy at recognizing the emotions of other. Neural Network is the one of the best algorithm for prediction and handling the voice and signal type of data. In the other types the emotion can be found using covert the files using the image processing but now we skipping the image processing using the algorithm MLP. Here we use only .Wav format files to represent an emotion of a human. Here we apply several techniques to find the emotion of the human. The first method will be a feature extraction and other will be prediction. Here we take eight type of emotions. They are sad, happy, calm, neutral, angry, fearful, disgust, surprised. Using that data we extract the emotion before the image processing model. In this we have two kind of modules which is test data and train data which will be useful in extraction of all the data that is used to fetch the emotion of the person.