Contactless seismocardiography via Gunnar-Farneback optical flow

Seismocardiography (SCG) has gained significant attention due to its potential applications in monitoring cardiac health and diagnosing cardiovascular conditions. Conventional SCG methods rely on accelerometers attached to the chest, which can be uncomfortable or inconvenient. In recent years, researchers have explored non-contact methods to capture SCG signals, and one promising approach involves analyzing video recordings of the chest. In this study, we investigate a vision-based method based on the Gunnar-Farneback optical flow to extract SCG signals from the chest skin movements recorded by a smartphone camera. We compared the SCG signals extracted from the chest videos of four healthy subjects with those obtained from accelerometers and our previous method based on sticker tracking. Our results demonstrated that the vision-based SCG signals extracted by the proposed method closely resembled those from accelerometers and stickers, although these signals were captured from slightly different locations. The mean squared error between the vision-based SCG signals and accelerometer-based signals was found to be within a reasonable range, especially between signals on head-to-foot direction (0.2$<$MSE$<$1.5). Additionally, heart rates derived from the vision-based SCG exhibited good agreement with the gold-standard ECG measurements, with a mean difference of 0.8 bpm. These results indicate the potential of this non-invasive method in health monitoring and diagnostics.

Reconstruction of 3-Axis Seismocardiogram from Right-to-left and Head-to-foot Components Using A Long Short-Term Memory Network

This pilot study aims to develop a deep learning model for predicting seismocardiogram (SCG) signals in the dorsoventral direction from the SCG signals in the right-to-left and head-to-foot directions ($\textrm{SCG}_x$ and $\textrm{SCG}_y$). The dataset used for the training and validation of the model was obtained from 15 healthy adult subjects. The SCG signals were recorded using tri-axial accelerometers placed on the chest of each subject. The signals were then segmented using electrocardiogram R waves, and the segments were downsampled, normalized, and centered around zero. The resulting dataset was used to train and validate a long short-term memory (LSTM) network with two layers and a dropout layer to prevent overfitting. The network took as input 100-time steps of $\textrm{SCG}_x$ and $\textrm{SCG}_y$, representing one cardiac cycle, and outputted a vector that mapped to the target variable being predicted. The results showed that the LSTM model had a mean square error of 0.09 between the predicted and actual SCG segments in the dorsoventral direction. The study demonstrates the potential of deep learning models for reconstructing 3-axis SCG signals using the data obtained from dual-axis accelerometers.