**Motivation** --- This dataset has been created for the [Pfizer Digital Medicine Challenge][1]. Early detection of respiratory tract infections can lead to timely diagnosis and treatment, which can result in better outcomes and reduce the likelihood of severe complications. Respiratory sounds carry rich information that can be mined to develop automated approaches for detection of sickness behaviors like coughing and sneezing. In this challenge, we invite you to build machine learning models for automatic detection of sickness sounds by using audio recordings from open datasets. The dataset was created using audio files from [ESC-50][2] and [AudioSet][3]. We used the open source [BMAT Annotation Tool][4] to annotate this dataset. **Challenge** --- Develop machine learning models for detection of sickness sounds (coughing and sneezing) **Dataset** --- The dataset is organized as follows: - audio / melspectrograms / spectrograms / continuous wavelet transform - train - sick (n=1435) - not_sick (n=2283) - validation - sick (n=468) - not_sick (n=753) - test - sick (n=642) - not_sick (n=1012) **Code Examples** --- **Python** - Transfer learning with Keras + TensorFlow (https://github.com/adanRivas/CNN-Audio-Classifier-with-Keras-Tensorflow) **MATLAB** - Coming soon **Other Resources:** --- **Audio Data Augmentation** - [Audio data augmentation methods in python – Kaggle][5] - [Kapre: Keras Audio Preprocessing Layers][6] - [Audio Preprocessing and Augmentation with Keras][7] **Keras** - [Useful Keras Features][8] - [Transfer Learning with Keras][9] - [Save and Load Keras Models][10] - [Hyperparameter Grid Search – Keras + Scikit-learn][11] - [Tensorboard with Keras][12] **Tensorflow** - [Vggish - pre-trained model trained on data from AudioSet][13] - [Save and Restore Tensorflow Models][14] - [Dataset Generator in Tensorflow][15] - [TensorBoard with Tensorflow][16] [1]: https://www.cmg.org/hack-pfizer/ [2]: https://github.com/karoldvl/ESC-50 [3]: https://research.google.com/audioset/ [4]: https://github.com/BlaiMelendezCatalan/BAT [5]: https://www.kaggle.com/CVxTz/audio-data-augmentation [6]: https://github.com/keunwoochoi/kapre#one-shot-example [7]: https://github.com/drscotthawley/panotti [8]: https://towardsdatascience.com/https-medium-com-manishchablani-useful-keras-features-4bac0724734c [9]: https://towardsdatascience.com/transfer-learning-using-keras-d804b2e04ef8 [10]: https://machinelearningmastery.com/save-load-keras-deep-learning-models/ [11]: https://machinelearningmastery.com/grid-search-hyperparameters-deep-learning-models-python-keras/ [12]: http://fizzylogic.nl/2017/05/08/monitor-progress-of-your-keras-based-neural-network-using-tensorboard/ [13]: https://github.com/tensorflow/models/tree/master/research/audioset [14]: http://cv-tricks.com/tensorflow-tutorial/save-restore-tensorflow-models-quick-complete-tutorial/ [15]: https://towardsdatascience.com/how-to-use-dataset-in-tensorflow-c758ef9e4428 [16]: https://www.tensorflow.org/programmers_guide/summaries_and_tensorboard