Indonesian Lipreading Dataset (IDLRW)

<div> <div>The first Indonesian Lip Reading Dataset (IDLRW), consists of 48,083 videos of 100 Indonesian words spoken by native speakers.</div> <div>The dataset is divided into 3 subsets with ratio (7:2:1) : training, validation, and testing.</div> <div>The videos resolution in 224x224 with 25 fps.</div> <div>The videos obtained from CC BY license videos that shared in online video sharing platform.</div> <div>&nbsp;</div> <div>For detailed information, please refer to the article "<strong>Recognizing Indonesian words based on visual cues of lip </strong><strong>movement using deep learning</strong>" published in the Measurement journal, Elsevier.</div> <div>Alternatively, you may contact us for more information about the dataset. (<a href="mailto:griffani.megiyanto@polban.ac.id">griffani.megiyanto@polban.ac.id</a>)</div> </div>