page_type	languages	description	urlFragment
sample & dataset	matlab	Exterior acoustic scattering data suitable for machine learning	acoustic-scattering-data

Acoustic Scattering Data

This repository contains synthetic acoustic scattering data for a random set of convex prism shapes represented as loudness fields in four octave bands, along with Matlab (TM) parsing and visualization sample scripts. The data was employed in the paper:

Ziqi Fan, Vibhav Vineet, Hannes Gamper, Nikunj Raghuvanshi,
Fast Acoustic Scattering using Convolutional Neural Networks,
IEEE ICASSP, 2020

Contents

For each type of dataset, there are folders for input binary image representing shape as occupancy grid and output as a four-channel image representing scattered spatial loudness maps in four octave bands, with pixels occupied by shape represented by NaN values. See the referenced paper for more details.

File/folder	Description
Train	Training data (zipped)
Val	Validation data
Test	Test data
Stretch	Generalization tests on analytic shapes
Visualization/visualizeData.m	Function illustrating parsing and visualizing the data. Takes two arguments: name of dataset and an array for indices of instances within the dataset
README.md	This README file.
CONTRIBUTING.md	Guidelines for contributing to the sample.

Prerequisites

Scripts were tested on Matlab v2017b.

Citing

If you employ the dataset, please cite using Bibtex key below.

@InProceedings{Fan_MLScattering:2020,
author = {Fan, Ziqi and Vineet, Vibhav and Gamper, Hannes and Raghuvanshi, Nikunj},
title = {Fast Acoustic Scattering using Convolutional Neural Networks},
booktitle = {IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)},
year = {2020},
month = {May},
}

License

The data and associated code are being released under the Open Use of Data Agreement, with the intention of promoting open research using this dataset.

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.