SiVi-CAFE dataset - Sighted and Visually-impaired Captions for Audio in Finnish and English

<p>This is a dataset containing <strong>audio captions</strong> for audio files of the <a href="../records/2589280">TAU Urban Acoustic Scenes 2019 development</a> dataset (airport, public square, and park) for 10 cities.&nbsp;</p> <p>The files were annotated using a web-based tool as presented in:</p> <p>Martin Morato, I., &amp; Mesaros, A. (2021). <a href="https://doi.org/10.5281/zenodo.5770113" target="_blank" rel="noopener"><em>Diversity and bias in audio captioning datasets</em></a>. In F. Font, A. Mesaros, D. P.W. Ellis, E. Fonseca, M. Fuentes, &amp; B. Elizalde (Eds.), Proceedings of the 6th Workshop on Detection and Classication of Acoustic Scenes and Events (DCASE 2021) (pp. 90-94)</p> <p>Each file is annotated by multiple annotators that provided a one-sentence description of the audio content.</p> <p><br>Data is provided in csv files:</p> <ul> <li>sighted-EN-bias-original</li> <li>sighted-FI-bias-translated</li> <li>sighted-EN-no_bias-original</li> <li>sighted-FI-no_bias-translated</li> <li>visually_impaired-FI-original</li> <li>visually_impaired-EN-translated</li> <li>sighted-FI-original</li> <li>sighted-EN-translated</li> </ul> <p>original = original descriptions, non-translated<br>translated = Translated descriptions using <a href="https://www.deepl.com/pro-api">automatic deep learning tool</a></p> <p>900 annotated audio files, Finnish audio descriptions provided by visual-impaired and sighted people.<br>2050 annotated audio files, English audio descriptions provided by international students (not-necessarily English native-speakers).<br>3930 annotated audio files, English audio descriptions provided by international students (not-necessarily English native-speakers) biased by the provided audio tags.</p> <p>&nbsp;</p> <p>The audio files can be downloaded from https://zenodo.org/record/2589280 and are covered by their own license.</p>