I. Project Overview: This project classifies sound signals from different environmental classes in the ESC-10 dataset. the above photo summarizes the model steps:
- The model read all the signals of different classes and assign a label number to each class.
- The Mel Frequency Cepstral Coefficients (MFCC) are extracted from the time domain.
- Full Convolutional Neural Network (CNN) is defined and used to classify 10 different classes of ESC-10 dataset
II. Dateset:
The dataset can be found in:
https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/YDEPUT