AudioFeatureVisualization

Feature Visualization applied to audio classification

This python script loads a pretrained VGG16 model on the GTZAN audio dataset and tries to visualize what the filters learned.

Sources :

• https://blog.keras.io/how-convolutional-neural-networks-see-the-world.html

Uses the following repos :

• https://github.com/Hguimaraes/gtzan.keras
• https://github.com/keras-team/keras/blob/master/examples/conv_filter_visualization.py