Deustek research team in collaboration with Zurich Ultrasound Research and Translation of the University Hospital Zurich and the Department of Radiology from Stanford University has published the article “Lung ultrasound for point-of-care COVID-19 pneumonia stratification: computer-aided diagnostics in a smartphone. First experiences classifying semiology from public datasets”.
In the paper, researchers propose the use of light-weight neural networks (MobileNets ) to support clinicians in the diagnostics and stratification of COVID-19, based on the extraction of LUS semiology. Preliminary developments and first experiences classifying common LUS semiology in COVID-19 patients using a customised neural network are shown. Researchers demonstrate that light-weight neural networks models, which can be readily deployed in mobile devices, can be used in computer-aided diagnostics of POCUS LUS-COVID-19 images. The initial promising results show that these models can provide competitive results in computationally constrained devices, making them ideal for their use in settings that require mobility and flexibility (i.e. patient stratification in nursing homes).
You can read the full article here.