Datasets
Standard Dataset
DDMP-FV
- Citation Author(s):
- Submitted by:
- si tong
- Last updated:
- Tue, 05/28/2024 - 11:28
- DOI:
- 10.21227/00tq-3495
- Data Format:
- License:
85 Views- Categories:
- Keywords:
Abstract
Health is a growing concern in modern society, and monitoring physiological indicators is an important part of maintaining health. Traditional health monitoring methods often require the use of contact sensors to monitor the human body, which is less convenient and comfortable, and often only measures relatively single physiological indicators, such as heart rate and blood oxygen. Traditional monitoring methods require complex instrumentation and sampling processes that require manual intervention, which is impractical for routine testing.
Facial video-based physiological indicator monitoring systems, on the other hand, can achieve the monitoring of various complex physiological indicators of the human body, such as respiratory rate, heart rate, oxygen saturation, blood pressure, etc., by sensing the external signals of the human body, and at the same time, they can also provide more comprehensive and fine monitoring solutions. Based on these more comprehensive and fine physiological indicator data, health analysis and diagnosis can be better carried out.
The dataset extracts heart rate and blood pressure through an electronic sphygmomanometer, heart rate and blood oxygen through a finger-clip oximeter, and manually records respiratory rate. This dataset provides an evaluation dataset for non-contact monitoring, and meets the needs of an evaluated, non-contact physiological indicator monitoring system.
We recruited 20 volunteers and before collecting the data, the experimenter pair showed each volunteer the objectives and the process of monitoring. The use of cameras and instruments was reasonable for each subject. Participants were given a consent form allowing them to choose whether or not to share their data with the scientific research community, which was signed by both the experimenter and the participant.
We created this dataset in order to evaluate the accuracy of physiological parameters extracted from facial videos. Dataset recorded the facial videos of 20 subjects in the resting, breath-holding resting, and locomotion resting states as well as the values corresponding to heart rate, blood oxygen, blood pressure, and respiratory rate. Each state was recorded twice, for 30 s each, for a total of 120 data sets. The device uses CONTEC CMS50DL pulse oximeter to record heart rate and blood oxygen; bioland arm-type electronic sphygmomanometer to record heart rate and blood pressure; and manual recording of respiratory rate.
