Data were collected from 142 subjects, yielding over 13,000 instances of synchronized data (3.8 TB).
Link to the published paper: https://www.mdpi.com/1424-8220/21/10/3465/htm
Data Acquisition
FLIR T540 thermal camera (464×348 pixels, 24◦ FOV) and a Logitech C920 Pro HD web-camera (768×512 pixels, 78◦ FOV) with a built-in dual stereo microphone were used for data collection purpose. Each subject participated in two trials where each trial consisted of two sessions. In the first session, subjects were silent and still, with the operator capturing the visual and thermal video streams through the procession of nine collection angles. The second session consisted of the subject reading a series of commands as presented one-by-one on the video screens, as the visual, thermal and audio data was collected from the same nine camera positions.
Citation
@Article{s21103465, AUTHOR = {Abdrakhmanova, Madina and Kuzdeuov, Askat and Jarju, Sheikh and Khassanov, Yerbolat and Lewis, Michael and Varol, Huseyin Atakan}, TITLE = {SpeakingFaces: A Large-Scale Multimodal Dataset of Voice Commands with Visual and Thermal Video Streams}, JOURNAL = {Sensors}, VOLUME = {21}, YEAR = {2021}, NUMBER = {10}, ARTICLE-NUMBER = {3465}, URL = {https://www.mdpi.com/1424-8220/21/10/3465}, ISSN = {1424-8220}, DOI = {10.3390/s21103465} }