This project presents an overview on the sound classification concept that consist in assigning a label to an unknown audio signal. This task is based on extracting relevant features from audio signals and using them to identify to which class the sound most likely belongs to.
For this aim, we first extracted Mel Frequency Cepstral Coefficients (MFCC) features from a sound data. After performing this step, we feed these features to a Convolutional Neural Network model (with Keras) to classify the audio signal into separate classes.
In the folder report you find the report about the project made explaining MFFC feature extraction details, explaining some details about Deep Learning and the realisation of the project.