基于预训练卷积神经网络的肺动脉压分类器研究

Research on Pulmonary Artery Pressure Classifier Based on Pre-Trained Convolutional Neural Network

肺动脉高压是一种肺血管系统的梗阻性持续性疾病,通常与先天性心脏病相关,可诱发右心室功能衰竭,影响患者的生活质量和生命。肺动脉压力(PAP)是判定肺动脉高压的直接指标,现有的肺动脉压力测量技术侵入性大,使用不便,不适合频繁使用。一种无创、方便和定期监测肺动脉压力的方法对于肺动脉高压的早期预防和诊断至关重要。本研究的目的是提出并评估一种深度学习方法,用于从无创的光电容积脉搏波(PPG)信号中分类和评估肺动脉高压。利用采集的肺动脉压力值提取肺动脉压力类别标签,PPG信号用于训练和测试模型。基于216条数据记录,使用预训练卷积神经网络进行分类,实验结果表明,该模型的分类准确率为97.78%,表明基于PPG信号提取特征训练的模型在PAP分类中具有良好的性能。随着可穿戴设备从指尖捕捉PPG信号的发展和深度学习模型的出现,无创和方便的PAP预测将极大地有助于心血管疾病的早期预防。

Pulmonary hypertension (PH) is an obstructive progressive disease of lung vascular system which is usually associated with congenital heart disease (CHD) and induces the right ventricle failure, adversely affecting quality of life and survival. Pulmonary arterial pressure (PAP) is a direct indicator of PH. The existing PAP measurement techniques are invasive and inconvenient, which are not suitable for frequent use. A noninvasive, convenient and regular monitoring PAP method is essential for the early prevention and diagnosis of PH. The aim of this study was to propose and evaluate a deep learning approach for the classification and evaluation of PH from noninvasive photoplethysmography (PPG) signals. The PAP signal was used to extract pulmonary artery pressure category label, the PPG signal was used to train and test model. A pre-trained convolutional neural network (GoogLeNet) was developed to learn features and classified PH based on the results from 216 data records, the experimental results showed that the classification accuracy of this model was 97.78%, which indicates that the GoogLeNet model trained on features extracted from PPG signals has excellent performance in PAP classification.