Datasets
Standard Dataset
The dataset of photoplethysmography signals collected from a pulse sensor to measure blood glucose level
- Citation Author(s):
- Submitted by:
- Ali Kermani
- Last updated:
- Fri, 12/13/2024 - 07:42
- DOI:
- 10.21227/mekr-0a81
- Data Format:
- License:
53 Views- Categories:
- Keywords:
Abstract
Non-invasive monitoring and surveillance methods of blood glucose measurement can provide ease of use and simplicity for different individuals while reducing the risks and damages of invasive methods. The non-invasive method based on photoplethysmography (PPG) signal is one of the innovative methods on this topic which numerous studies have been conducted by research centers and various companies. However, due to various reasons, the reviewed dataset was not available and no standard dataset has been published on this topic. The presented dataset, which was sampled by the research team of the digital systems of the University of Science and Technology in Mazandaran, Behshahr, Iran, includes 67 raw PPG signals with 2175 Hz sampling frequency along with labeled data: age, gender, and the invasive blood glucose level which can be used in other studies and learning algorithms.
In collecting the dataset, a total of 67 PPG signal samples were obtained from 23 individuals. The sampling process was carried out as follows: after completing a questionnaire by the volunteers, the PPG signal was obtained from the first band of the index finger using the pulse sensor. Then, blood glucose levels were measured using the invasive method with the Accu-CheckTM [1] self-monitoring device of the Active series with a calibration kit of 333.the labeling was performed respectively. The pulse sensor used in this study employs a green-colored 550-nm optical transmitter and an APDS9008 photodiode receiver. The received signal is amplified by a low-frequency operational amplifier (MCP6001) and passed through an RC filter to control the amplification of the signal and the noise carried along with it. The carrier signal is then passed through a rectifier diode (inverse leakage current protector) and enters the ADC pin of the Atmega328TM microcontroller. In the process of obtaining the PPG signal from individuals, factors such as avoiding movement were considered to ensure that the obtained signal meets the conditions of a standard PPG signal to use in the studies. This dataset includes 67 samples of PPG signals obtained from the fingertip area of individuals. These signals were sampled from different individuals under appropriate conditions using a pulse sensor consists of APDS-9008 device The dataset consists in an archive file named ‘PPG_Dataset.zip’, containing three folders: ‘RawData’ , ‘Labels’, and 'Figures' which contain 67 mat files, and 68 mat files, and 67 Matlab figure files ,respectively. The raw data mat files (in RawData folder) contain PPG signals correspond to the 12 bits ADC output of the photodetector lighted with the green LED. It was named as 'signal_XX_XXXX.mat' in which XX and XXXX are corresponded to the partipant ID and to the number of data acquized from one participant. The total duration of recording session is 674621 milliseconds. The 'Labels' folder contains 'Label_XX_XXXX.mat' files named as similar as the raw data file name. Each label contains a table whose first column is the participant ID (1-23), second column is gender(male/female), third column is age (years), forth column is blood glucose level (mg/dLitr), fifth column is height, and sixth column is weight for each participant, totally 23 volunteers. The 'Figures' folder contains the images of signals in JPG format : 'fig_XX_XXXX.jpg'. The naming is the same as the raw data file name.
