IMVIA-NIR

Overview

IMVIA-NIR is a new, publicly available dataset designed for research on remote photoplethysmography (rPPG). It consists of near-infrared videos collected from a diverse group of 10 subjects in an indoor setting. The dataset is divided into two subsets: 'still', where subjects were static, and 'talking', where subjects were speaking (20 videos in total). This dataset addresses the scarcity of public resources for rPPG research, particularly in the realm of NIR vision. 

For more details, please refer to the BSPC'24 paper.

Materials

• Camera: IDS UI-3240ML-NIR-GL

• Videos saved at 20 fps in MJPEG format, 1280*1024 pixels

• Light Source: NIR LED 850nm (with tracing paper diffuser)

• Reference Device: Empatica E4 watch (BVP.csv files, 64 Hz)

Notes

• Camera and E4 watch synchronized manually (potential for minor misalignment)

• Some videos may exhibit saturation, sorry for that

Contact

Yannick Benezeth (yannick.benezeth@u-bourgogne.fr)

Download

You can download the dataset here. 

Citation

If you use the IMVIA-NIR dataset in your work, please cite:

Benezeth, Y., Krishnamoorthy, D., Monsalve, D. J. B., Nakamura, K., Gomez, R., & Mitéran, J. (2024). Video-based heart rate estimation from challenging scenarios using synthetic video generation. Biomedical Signal Processing and Control, Elsevier.