A project from Digital Signal Processing course
- Python 2.7 (3.6 not tested)
- numpy
- librosa
- pysoundfile
- matplotlib
- scikit-learn
- tensorflow
- keras
Dataset could be downloaded at Dataverse or Github.
I'd recommend use ESC-10 for the sake of convenience.
Put audio files under data directory and run the following command:
$ python feat_extract.pyFeatures and labels will be generated and saved in the directory.
Make sure you have scikit-learn installed and feat.npy and label.npy under the same directory. Run svm.py and you could see the result.
Install tensorflow and keras at first. Run nn.py to train and test the network.
Run cnn.py to train and test a CNN.