Abstract:Hypertension is a medical condition characterized by high blood pressure, and classifying it into its various stages is crucial to managing the disease. In this project, a novel method is proposed for classifying stages of hypertension using Photoplethysmography (PPG) signals and deep learning models, namely AvgPool_VGG-16. The PPG signal is a non-invasive method of measuring blood pressure through the use of light sensors that measure the changes in blood volume in the microvasculature of tissues. PPG images from the publicly available blood pressure classification dataset were used to train the model. Multiclass classification for various PPG stages were done. The results show the proposed method achieves high accuracy in classifying hypertension stages, demonstrating the potential of PPG signals and deep learning models in hypertension diagnosis and management.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to use photoplethysmography (PPG) signals and deep - learning models to classify different stages of hypertension. Specifically, the authors propose a new method that uses PPG signals combined with a deep - learning model (especially the AvgPool_VGG - 16 model) to identify and classify different stages of hypertension, thereby improving the accuracy of hypertension diagnosis and management. ### Background and Objectives of the Paper Hypertension is a medical condition mainly characterized by elevated blood pressure and can be divided into different stages according to its severity. Accurately classifying these stages is crucial for the management and treatment of the disease. Traditional blood pressure measurement methods usually require the use of a cuff. This method is not only invasive but also not easy to be carried out frequently in daily life. Therefore, developing a non - invasive and convenient method for classifying hypertension has important clinical value. ### Solutions The authors propose a new method based on PPG signals and deep - learning models. PPG signals are a method of measuring changes in blood volume through optical sensors and can be used to indirectly estimate blood pressure. The deep - learning models used in the paper include: - **AlexNet**: A classic convolutional neural network model used for image classification tasks. - **ResNet - 50**: A deep residual network that can effectively alleviate the vanishing gradient problem in deep networks. - **VGG - 16**: A convolutional neural network with 13 convolutional layers and 3 fully - connected layers, widely used in image classification. - **AvgPool_VGG - 16**: The improved model proposed in this paper, which uses average pooling instead of max pooling to better preserve temporal information and reduce noise. ### Experimental Results By classifying the PPG signals of 219 patients, the experimental results show that: - **AlexNet**: Performed the worst and was unable to correctly classify any signals. - **ResNet - 50**: Performed better than AlexNet but still had some classification errors. - **VGG - 16**: Significantly outperformed other models, with an accuracy rate of 71% and an F1 - score of 0.69. - **AvgPool_VGG - 16**: Performed the best, with an accuracy rate of 80% and an F1 - score of 0.77. ### Conclusions The method proposed in the paper performs excellently in classifying different stages of hypertension, especially when using the average pooling technique. This indicates that PPG signals combined with deep - learning models, especially models using average pooling, can play an important role in hypertension diagnosis. Future research can further increase the scale of the data set, explore more signal - processing methods, and try different pooling techniques to further improve the accuracy and robustness of classification.

PPG Signals for Hypertension Diagnosis: A Novel Method using Deep Learning Models