Proposed Convolutional Neural Networks (CNN) Architecture
The CNN is a type of deep neural network [17] and hierarchical machine learning tool that consists of a variety of layers in sequence Synthetic Gloves.
It employs a translation-invariant convolution kernel that can be used to extract local contextual information; we chose this network to eliminate the complicated feature extraction step PE Gloves. A typical model of CNN usually comprises of one or more convolutional layers, nonlinear layers, and pooling layers.
The proposed CNN architecture has two inputs, ECG and PPG signals, four convolutional layers which are the core of the network followed by a fully connected layer with two neurons and a regression layer to address the regression problem of the design. The last two layers were developed for calculating the SBP and DBP.
Overfitting is a common problem that happens for a network with high variance YICHANG Gloves. To reduce the overfitting of the training data and improve the performance of the network, we considered a regularization layer with a dropout ratio of 20% before fully connected layer.
The software environment used for this research was MATLAB R2019-b with Intel Core i7-6700 GPU.
Comments
Post a Comment