Deep Neural Network topologies for audio classification. This repo includes a variety of topologies, including:
Dense, Dense-LSTM,
1D-CNN, 1D-CNN-LSTM,
2D-CNN, 2D-CNN-LSTM
The implementations are powered by Python.
Specific implementations are described in the following papers:
"1D & 2D CNNs vs Temporal Aggregated Feature-Based Methodologies for Audio Classification"
"Continuous Speech Emotion Recognition with Convolutional Neural Networks"
You can experiment using the LVLib or BDLib Datasets:
http://research.playcompass.com/files/LVLib-SMO-1.zip
http://research.playcompass.com/files/LVLib-SMO-2.zip
http://research.playcompass.com/files/LVLib-SMO-3.zip
http://research.playcompass.com/files/BDLib-1.zip
http://research.playcompass.com/files/BDLib-2.zip