DCASE2020-Anomly-Detection

This repo is actually taken from "https://github.com/parth2170/DCASE2020-Task2" and the issues related to the library versions are resolved.

This repository can be used to reproduce our submissions for DCASE Challenge 2020 Task 2 - Unsupervised Detection of Anomalous Sounds for Machine Condition Monitoring

Abstract

This report summarizes our submission for Task-2 of the DCASE 2020 Challenge. We propose two different anomalous sound detection systems, one based on features extracted from a modula- tion spectral signal representation and the other based on i-vectors extracted from mel-band features. The first system uses a nearest neighbour graph to construct clusters which capture local variations in the training data. Anomalies are then identified based on their distance from the cluster centroids. The second system uses i-vectors extracted from mel-band spectra for training a Gaussian Mixture Model. Anomalies are then identified using their negative log likelihood. Both these methods show significant improvement over the DCASE Challenge baseline AUC scores, with an average improvement of 6% across all machines. An ensemble of the two systems is shown to further improve the average performance by 11% over the baseline.

Requirements

conda create -n <env-name> python==3.8
conda activate <env-name>
git clone https://github.com/jfsantos/SRMRpy.git
cd SRMRpy/
python setup.py install
pip install -r requirements.txt

Usage

1. Clone this repository

2. Download datasets

• Datasets are available here
• Datasets for all machines can be downloaded and unzipped by running
  • sh download_dev_data.sh for development data
  • sh download_eval_data.sh for evaluation data

3. Running System 1

• cd bin/modspec_graph/
• python graph_anom_detection.py d - for running on development data
  • Modulation Spectrums for each machine-id will be stored in npy files in saved/ in the same directory
  • The results for development data are stored in modspec_graph_dev_data_results.csv in the same directory
• python graph_anom_detection.py e - for running on evaluation data
  • The results for evaluation data are stored in the submission format in the directory task2

3. Running System 2

• i-Vectors for both development and evaluation have been provided in the zip file - saved_iVectors/ivector_mfcc_100.zip
• Unzip ivector_mfcc_100.zip in the same directory
  • Code for extracting i-Vectors will be added soon
• cd bin/iVectors_gmm/
• python gmm.py d - for running on development data
  • The results for development data are stored in iVectors_gmm_dev_data_results.csv in the same directory
• python gmm.py e - for running on evaluation data
  • The results for evaluation data are stored in the submission format in the directory task2