The HumBug Challenge: ComParE 2022

<p><strong>A large-scale multi-species dataset of acoustic recordings</strong></p> <p>Dataset compatible with two papers:</p> <ul> <li>The&nbsp;<strong>Computational Paralinguistics ChallengE (ComParE): Mosquito Event Detection Task</strong><strong>&nbsp;</strong><a href="https://github.com/EIHW/ComParE2022/tree/MOS-C">https://github.com/EIHW/ComParE2022/tree/MOS-C</a></li> <li>An update to:&nbsp;<em>HumBugDB: a large-scale acoustic mosquito dataset:</em> <ul> <li><a href="https://arxiv.org/abs/2110.07607">NeurIPS 2021 Paper</a></li> <li><a href="https://github.com/HumBug-Mosquito/HumBugDB">https://github.com/HumBug-Mosquito/HumBugDB</a>.</li> </ul> </li> </ul> <p>A large-scale multi-species dataset containing recordings of mosquitoes collected from multiple locations globally, as well as via different collection methods.&nbsp;In total, we present&nbsp;20 hours&nbsp;of labelled mosquito data with 15 hours&nbsp;of corresponding background noise, recorded at the sites of 8 experiments.&nbsp;&nbsp;Of these, 64,843 seconds contain species metadata, consisting of 36 species (or species complexes).</p> <p>This repository contains:</p> <ul> <li>Audio files to be extracted into <em>audio/data/</em><em>train</em>&nbsp;and&nbsp;<em>audio/data/</em><em>dev/{a/b} respectively</em></li> <li>Metadata in <em>csv</em> format:&nbsp;<a href="https://zenodo.org/record/6478589/files/humbugdb_zenodo_0_0_2.csv?download=1">neurips_2021_zenodo_0_0_2.csv</a></li> </ul> <p>&nbsp;</p>