Sound event localization and detection (SELDnet) results

<p>This package is part of the work -&nbsp;<a href="https://github.com/sharathadavanne/seld-metric">Joint Measurement of Localization and Detection of Sound Events</a>&nbsp;presented in WASPAA 2019.</p> <p>This package consists of results from the <a href="https://arxiv.org/abs/1905.08546">SELDnet method</a> for joint&nbsp;sound event localization and detection. The results corresponding to different training states of 5, 25 and 75 epochs are provided here.&nbsp; These results are for the four cross-validation splits of the&nbsp;<strong>TAU Spatial Sound Events 2019 - Microphone Array </strong>dataset. The sound events in this dataset&nbsp;consist of stationary point sources from multiple sound classes each associated with a temporal onset and offset time, and DOA coordinate represented using azimuth and elevation angle.&nbsp;This <strong>TAU Spatial Sound Events 2019 - Microphone Array </strong>dataset is part of the&nbsp;<a href="https://github.com/sharathadavanne/seld-dcase2019">DCASE 2019 Sound Event Localization and Detection Task</a>&nbsp;and can be downloaded <a href="https://zenodo.org/record/2599196#.XT_RmHUzaCg">here</a>.</p> <p>Each of the results folders consists&nbsp;of 400 files, corresponding to the results of the individual recordings of the&nbsp;<strong>TAU Spatial Sound Events 2019 - Microphone Array </strong>dataset.</p> <p>This data collection received funding from the European Research Council, grant agreement 637422 EVERYSOUND.</p> <p><strong>Download instructions</strong></p> <p>The three files, &nbsp;<strong><em>mic_dev_5</em></strong>,<strong><em>&nbsp;mic_dev_25,&nbsp;</em></strong>and&nbsp;<strong><em>mic_dev_75</em></strong>, correspond to SELDnet results at 5, 25 and 75 epochs.</p> <p>Download the zip files&nbsp;and use your favorite compression tool to unzip these split zip files.</p>