Description:According to the activity diagram (Fig.1), the user initially provides a voice input file via the web app. If the user provides the proper voice input file, he can request emotion from the web program by clicking the Make Prediction button. If the user's input file is incorrect or he or she has picked an unsupported voice file, the user must select another supported and appropriate voice file. After providing the relevant input file, the user can ask the web app for the emotion of that specific voice input file. By communicating with the system, the web app stores the supplied input voice file in the system, extracts features from it, and obtains the saved model path. Using these selected features and the saved model, the web app maps the selected features to a certain emotion by using the saved model to make a prediction, and then displays the outcome for the provided input voice file and Plays the music playlist as per the emotion detected.
, Claims:We Claim
1. A method for classifying Music based on the Speech Emotion Detected comprising of Librosa and PyAudio Packages from Python, Audio files from RAVDESS Dataset using MLP Classifier to test and train the ML Model.
2. The method of claim 1 wherein data is retrieved from the RAVDESS dataset.
3. The method of claim 1 further comprising preliminary analysis from RAVDESS dataset by preprocessing it using standardization method.
4. The method of claim 1 wherein Multilayer Perceptron Classifier (MLP) is used to train, test and classify the data from speech signals.
5. The method of claim 1 where we extracted MFCC, Chroma, MEL features from the audio file as the Descriptive features to detect the emotion from the audio file.