13 January 2024 in ACM Transactions on Computing for Healthcare
Inter-beat interval (IBI) measurement enables estimation of heart-tare variability (HRV) which, in turn, can provide early indication of potential cardiovascular diseases (CVDs). However, extracting IBIs from noisy signals is challenging since the morphology of the signal gets distorted in the presence of noise. Electrocardiogram (ECG) of a person in heavy motion is highly corrupted with noise, known as motion-artifact, and IBI extracted from it is inaccurate. As a part of remote health monitoring and wearable system development, denoising ECG signals and estimating IBIs correctly from them have become an emerging topic among signal-processing researchers. Apart from conventional methods, deep-learning techniques have been successfully used in signal denoising recently, and diagnosis process has become easier, leading to accuracy levels that were previously unachievable. We propose a deep-learning approach leveraging tiramisu autoencoder model to suppress motion-artifact noise and make the R-peaks of the ECG signal prominent even in the presence of high-intensity motion. After denoising, IBIs are estimated more accurately expediting diagnosis tasks. Results illustrate that our method enables IBI estimation from noisy ECG signals with SNR up to -30 dB with average root mean square error (RMSE) of 13 milliseconds for estimated IBIs. At this noise level, our error percentage remains below 8% and outperforms other state-of-the-art techniques.
17 May 2021 in IEEE Sensors Journal
DWDiane Myung-Kyung Woodbridge
Diane Myung-Kyung Woodbridge
+1
Fluid intake tracking is crucial in providing interventions that assist individuals to stay hydrated by maintaining an adequate amount of fluid. It also helps to manage calorie intake by accounting for the amount of calorie consumed from beverages. While staying hydrated and controlling calorie intake is critical in both physical wellness and cognitive health, existing technologies do not provide a solution for monitoring both fluid type and fluid volume. To address this limitation, we present the design, implementation, and validation of Liquid Intake Detection System (LIDS) for real-time tracking of fluid intake type and volume. The system devises a sensing module that is composed of ultrasonic, RGB color, temperature, and accelerometer sensors as well as a computational framework for machine-learning-based fluid intake type classification, volume estimation, and bottle-state-recognition. The developed sensing unit is small and light-weight that can be mounted, from inside, on the lid of a drinking bottle. We conduct extensive experiments to collect data in a variety of bottles and environmental settings. The results show that the accuracy of fluid type detection ranges from 74.93% to 94.98% while trying to detect the fluid of an unseen bottle. Our results for volume estimation show that the regression-based volume estimation supports a root-relative-squared-error that ranges from 1.12% to 13.36%.