IPPG Signals and Corresponding Blood Pressure Data Extracted from CN-BP and VV-Small Datasets

CN-BP is the publicly available CN-BP dataset in our previous work. It contains facial videos and blood pressure data of 43 Asians. The videos have a resolution of 1920x1080 and a frame rate of 30 FPS. The VV-Small dataset was published by Toye Pieter-Jan in 2023. We removed some data in which facial recognition was not feasible due to overexposure, low lighting, and other factors. The remaining facial videos and blood pressure data of 85 different volunteers were used as experimental data. The links to these two datasets are provided below in "Related links".

"CN-BP_IPPG.npy" and "VV-Small_IPPG.npy" store the IPPG signals extracted from facial videos respectively. The signals of each tested person are segmented into 3 signals with a length of 250. "CN-BP_DBP.npy" and "CN-BP_SBP.npy" store the DBP (mmHg) and SBP (mmHg) data of each tested person respectively, and the storage order is consistent with that of the IPPG signals. That is to say, each tested person has 3 signals with a length of 250, and at the same time, corresponds to one DBP and one SBP data in order. The same principle applies to the VV-Small dataset.

CN-BP是我们此前研究中公开发布的CN-BP数据集，包含43名亚洲人的面部视频与血压数据。该数据集的视频分辨率为1920×1080，帧率为30 FPS。VV-Small数据集由Toye Pieter-Jan于2023年发布。我们剔除了因过曝、低光照等因素导致无法完成面部识别的数据，剩余85名不同志愿者的面部视频与血压数据被用作实验数据。两个数据集的链接已在下文"相关链接"中提供。"CN-BP_IPPG.npy"与"VV-Small_IPPG.npy"分别存储了从面部视频中提取的成像光电容积脉搏波（IPPG）信号。每名受试者的IPPG信号被分割为3段长度为250的子信号。"CN-BP_DBP.npy"与"CN-BP_SBP.npy"分别存储了每名受试者的舒张压（DBP，mmHg）与收缩压（SBP，mmHg）数据，其存储顺序与IPPG信号的存储顺序一致。换言之，每名受试者对应3段长度为250的信号，同时按顺序对应一组舒张压与收缩压数据。VV-Small数据集亦遵循相同规则。

搜集汇总

背景与挑战

背景概述

该数据集整合了来自CN-BP和VV-Small两个公开数据集的面部视频提取的IPPG信号及其对应的血压数据，其中CN-BP包含43名亚洲人数据，VV-Small包含85名志愿者数据。数据集提供每个受试者的IPPG信号（分段为3个长度为250的信号）和相应的舒张压（DBP）与收缩压（SBP）值，存储顺序一致，便于关联分析。该数据集适用于非接触式血压监测和生理信号处理研究，支持基于计算机视觉的血压估计方法开发。