A Multiwavelength Photoplethysmography Dataset with Blood Pressure and Heart Rate Reference Measurements

This dataset contains physiological recordings and reference values from 127 subjects, aimed at supporting research on blood pressure estimation using photoplethysmographic (PPG) signals. For each subject, the file input_data.mat provides pre-processed PPG signals stored in a 127×1 MATLAB cell array. Each cell contains a 1000×2 double matrix, where the first column represents the infrared PPG signal and the second column represents the red PPG signal. These signals have been standardized, band-pass filtered between 0.5 and 10 Hz, cleaned of spikes using a Hampel filter, and linearly interpolated to remove missing values. The file output_data.mat includes the reference blood pressure values in a 127×2 uint8 matrix. The first column contains systolic blood pressure measurements, and the second column contains diastolic blood pressure measurements, both in mmHg. Additionally, physiological_data.mat provides supplementary demographic and physiological information in a 127×3 uint8 matrix. This matrix includes each subject’s age, sex (encoded as 1 for male and 0 for female), and heart rate in beats per minute (bpm).
The dataset is suitable for developing and validating machine learning models for non-invasive blood pressure estimation.