README.md

Convolutional Neural Net to distinguish Foosball sounds from other noises

This is a CNN that distinguishes between the sounds of foosball goals and non-goals. Included is a pickled dataset ready for use, and also the original sound recordings and foosball goals timestamps if you want to try something different.

IPython notebook: CNN on Foosball sounds.ipynb

Trained CNN model using TensorFlow (version 0.8.0): model.ckpt

Pickled Pandas dataframe (version 0.18.0): full_dataset_44100.pickle

Detailed blog post: https://humblesoftwaredev.wordpress.com/2016/05/02/an-audio-dataset-and-ipython-notebook-for-training-a-convolutional-neural-network-to-distinguish-the-sound-of-foosball-goals-from-other-noises-using-tensorflow/