Description
This dataset includes a variety of motion scenarios designed to capture different types of movements and expressions, recorded using high-quality RGB cameras and synchronized with physiological signals.
Dataset Details
Subjects: 33 subjects (26 male, 7 female) with age range 21 to 25 years
Camera: RGB Logitech C920 camera - 30fps, 640x480 pixels stored in an uncompressed YUV planar pixel format (approx. 1GB per video)
Reference signal: Contec CMS50E pulse oximeter - 60Hz PPG signal
Lighting setup: 1) Natural light coming from a nearby window. 2) Two 155W LED with a color temperature of 3200K
Activities:
- Speaking: Participants were seated in front of a camera. They were given cards with scrambled words to read aloud, involving head movements and facial expression changes.
- Anterior-posterior Movement: Participants performed forward and backward movements, changing the distance between their faces and the camera (at least 20 cm, 45 movements per minute).
- Horizontal Movement Scenario: Participants performed horizontal body movements (at least 45 movements per minute).
- Random Rotation: Participants performed random head rotations.
Download
Download the request form, print it, sign it, and send it to rcsong@hfut.edu.cn. Once received, we will share a download link. The dataset size is approximately 272 GB. It is distributed in a single file compressed with a 7zip program.
Manual
AVI/
||---XX/
||-------YY/
||------------ZZ/
||---------------video.avi
||---------------wave.csv
where XX is the subject ID, YY is the light ID, ZZ is the activity ID, and file is one of {"video.avi", "wave.csv"}.
XX = {00,01,...,16}.
YY - lighting setup
0 - Natural light
1 - LED lighting
ZZ - activity
0 - Speaking
1 - Anterior-posterior Movement
2 - Horizontal Movement
3 - Random Rotation
Citation
If you use this dataset in your research, please cite the following publication: