<b>A Multi-Pathology Ballistocardiogram Dataset for Cardiac Function Monitoring and Arrhythmia Assessment</b>
收藏资源简介：
This dataset comprises 153 subfolders within a primary directory named <i>data</i>, derived from 85 participants. Each participant typically contributes 2–3 subfolders, contingent on the completeness and quality of their M-mode echocardiography (UCG) recordings. Subfolder names follow the format: hdata + SubjectID + EJ/XJ/ZJ to denote the specific cardiac region captured in the ultrasound data:EJ denotes M-mode imaging of the mitral valve, XJ denotes M-mode imaging of the left ventricle, and ZJ denotes M-mode imaging of the aortic valve.For instance, a participant with identifier “001” may have subfolders named hdata1EJ, hdata1XJ, and/or hdata1ZJ, corresponding to each available M-mode echocardiographic segment. Each subfolder contains five distinct files, described in detail below.<b>1 BCG J-peak file</b>(1) File name: hdata+subjectID+EJ/XJ/ZJ_BCG.csv(2) Content: J-peak positions in the BCG signal, presented in two columns:(3) The first column provides the raw data point index.(4) The second column specifies the corresponding time (in seconds) for each J-peak.<b>2 ECG R-peak file</b>(1) File name: hdata+subjectID+EJ/XJ/ZJ_ECG.csv(2) Content: R-peak positions in the ECG signal, also in two columns:(3) The first column provides the raw data point index.(4) The second column specifies the corresponding time (in seconds) for each R-peak.<b>3 Ultrasound video</b>(1) File name: hdata+subjectID+EJ/XJ/ZJ_UCG.AVI(2) Content: An AVI-format video of the simultaneously acquired M-mode echocardiogram. The suffix EJ, XJ, or ZJ indicates whether the imaging targeted the mitral valve, left ventricle, or aortic valve, respectively.<b>4 Signal data</b>(1) File name: signal.csv(2) Content: Three columns of time-series data sampled at 100 Hz. Raw BCG signal (Column 1).ECG data (Lead V2 or another designated lead) (Column 2). Denoised BCG signal (Column 3), derived using the Enhanced Singular Value Thresholding (ESVT) algorithm.<b>5 Signal visualization</b>(1) File name: signal.pdf(2) Content: A graphical representation of the signals from signal.csv. This file facilitates quick inspection of waveform alignment and overall signal quality.In addition to the data directory, an Additional_info folder provides participant demographic and clinical details. Each row in subject_info.csv corresponds to an individual participant, listing their ID, sex, weight, height, age, heart rate, ejection fraction(EF) (%). These parameters establish an informative link between each participant’s anthropometric profile, cardiac function metrics, and the corresponding BCG, ECG, and ultrasound data.<br>
本数据集包含153个子文件夹，均置于名为<i>data</i>的主目录下，数据源自85名受试者。每名受试者通常贡献2~3个子文件夹，具体数量取决于其M型超声心动图（M-mode echocardiography, UCG）录制文件的完整性与质量。子文件夹命名格式为：hdata + 受试者编号 + EJ/XJ/ZJ，用于标识超声数据采集的特定心脏区域：其中EJ代表二尖瓣M型成像，XJ代表左心室M型成像，ZJ代表主动脉瓣M型成像。例如，编号为“001”的受试者可能拥有名为hdata1EJ、hdata1XJ和/或hdata1ZJ的子文件夹，分别对应其可用的各M型超声心动图节段。每个子文件夹均包含5个不同文件，详述如下。 <b>1 心冲击图J峰文件（BCG J-peak file）</b> (1) 文件名格式：hdata+受试者编号+EJ/XJ/ZJ_BCG.csv (2) 文件内容：心冲击图（Ballistocardiography, BCG）信号中的J峰位置，以两列形式呈现： (3) 第一列为原始数据点索引； (4) 第二列为各J峰对应的时间（单位：秒）。 <b>2 心电图R峰文件（ECG R-peak file）</b> (1) 文件名格式：hdata+受试者编号+EJ/XJ/ZJ_ECG.csv (2) 文件内容：心电图（Electrocardiography, ECG）信号中的R峰位置，同样以两列形式呈现： (3) 第一列为原始数据点索引； (4) 第二列为各R峰对应的时间（单位：秒）。 <b>3 超声视频文件（Ultrasound video）</b> (1) 文件名格式：hdata+受试者编号+EJ/XJ/ZJ_UCG.AVI (2) 文件内容：同步采集的M型超声心动图AVI格式视频。后缀EJ、XJ、ZJ分别标识成像目标为二尖瓣、左心室或主动脉瓣。 <b>4 信号数据文件（Signal data）</b> (1) 文件名：signal.csv (2) 文件内容：3列100Hz采样的时序数据：第1列为原始心冲击图（BCG）信号；第2列为心电图（ECG）数据（V2导联或其他指定导联）；第3列为经增强奇异值阈值化（Enhanced Singular Value Thresholding, ESVT）算法处理得到的去噪后心冲击图信号。 <b>5 信号可视化文件（Signal visualization）</b> (1) 文件名：signal.pdf (2) 文件内容：signal.csv中信号的图形化展示，用于快速检查波形对齐情况与整体信号质量。 除上述数据目录外，另有一个Additional_info文件夹，提供受试者的人口学与临床细节信息。subject_info.csv中的每一行对应一名受试者，记录了其编号、性别、体重、身高、年龄、心率及射血分数（ejection fraction, EF，单位：%）。这些参数可建立受试者人体测量学特征、心脏功能指标与对应的心冲击图、心电图及超声数据之间的关联。
