This repository is an implementation of Noise Masking RNN for respiratory sound classification proposed in our paper

Video presentation about the project

It includes a preprocessing functions to convert raw .wav respiratory sound files and train script

Download model and data dirs from here

Run python3 train.py --gpu GPU_NUM --data_path DATA_PATH --cv_path CV_SPLIT_PATH --exp_path EXPERIMENT_PATH

(data and cv split are included here in /data directory)

Run python3 predict.py --wav WAVFILE_PATH

Script prints text: "Probability of anomalies: X%"