An approach to a wrist wearable based Covid-19 prediction system to protect Health Care Professionals.

With healthcare professionals being the frontline warriors in battling the Covid pandemic, their risk of exposure to the virus is extremely high. We present our experience in building a system, aimed at monitoring the physiology of these professionals 24/7, with the hope of providing timely detection of infection and thereby better care. We discuss a machine learning approach and model using signals from a wrist wearable device to predict infection at a very early stage.

Ensembled Transfer Learning and Multiple Kernel Learning for Phonocardiogram Based Atherosclerotic Coronary Artery Disease Detection.

Conventional machine learning has paved the way for a simple, affordable, non-invasive approach for Coronary artery disease (CAD) detection using phonocardiogram (PCG). It leaves a scope to explore improvement of performance metrics by fusion of learned representations from deep learning. In this study, we propose a novel, multiple kernel learning (MKL) for their fusion using deep embeddings transferred from pre-trained convolutional neural network (CNN).

Novel features from autocorrelation and spectrum to classify Phonocardiogram quality.

Phonocardiogram (PCG) or auscultation via a stethoscope forms the basis of preliminary medical screening. But PCG recorded in an uncontrolled environment is inherently noisy. In this paper we have derived novel features from the spectral domain and autocorrelation waveforms.

Better anticoagulation control improves survival after valve replacement.

Objective: We sought to assess the effect of anticoagulation control on long-term survival after valve replacement with the Medtronic Hall valve (Medtronic, Inc, Minneapolis, Minn).

Methods: Prospective follow-up data, including 82,297 international normalized ratios, were collected for 1476 patients undergoing single valve replacement with the Medtronic Hall valve between 1979 and 1994, with follow-up to the end of 1998. After excluding 204 patients who either died within 30 days or had fewer than 10 international normalized ratios recorded beyond 30 days, there were 10,203 patient years of follow-up for analysis.