The COVID-19 virus induces infection in both the upper respiratory tract and the lungs. Chest X-ray are widely used to diagnose various lung diseases. Considering chest X-ray and CT images, we explore deep-learning-based models namely: AlexNet, VGG16, VGG19, Resnet50, and Resnet101v2 to classify images representing COVID-19 infection and normal health situation. We analyze and present the impact of transfer learning, normalization, resizing, augmentation, and shuffling on the performance of these models. We explored the vision transformer (ViT) model to classify the CXR images. The ViT model incorporates multi-headed attention to disclose more global information in constrast to CNN models at lower layers. This mechanism leads to quantitatively diverse features. The ViT model renders consolidated intermediate representations considering the training data. For experimental analysis, we use two standard datasets and exploit performance metrics: accuracy, precision, recall, and F1-score. The ViT model, driven by self-attention mechanism and longrange context learning, outperforms other models.
Given the pandemic infection risk in classrooms and given the potential to purify COVID-19 prone air, this research team has visualized the flow of air to find the optimal position in a room. Through Schlieren imaging the air flow was studied to establish the circulation in the tested room. With a variation of air purifier positions in a model classroom, the imaging sensors have taken profiles of airflow and therefore contributed to identifying the optimal placings in heated classrooms. Given a random position of a potentially infected and COVID-19 infectious person, the systematic research measured concentrations of artificially produced particles that emulated aerosol distributions. The research established contaminations stabilizing after a quarter of an hour. The concentrations are only a fraction of the emitted effluents. In this way, the risk of superspreading can be mitigated and so the results allow continued academic work during the Corona pandemic
Increasing COVID-19 infections are reason of concern for all the inside workplaces where physical presence is necessary for collaborating. Classrooms are one of the suspected places, where usually students are closely placed to learn together as in times before the pandemic. To reduce the infection rate in classrooms, an air purifier was designed around a commercial filter which removes 99,9% of particles with 3μm. A baseline optical study of air purification was carried out to ensure effectiveness of the purifier during operation in closed environment. With conclusive evidence of microscopic images, breathing tests and aerosol penetration test using oil, the filter effectiveness was recorded. Optical values for suspended particle counts are recorded for variations in air flow rates of the air purifier and the gradual change is helping to understand the filter performance. Already around 70% minimum effectiveness of one flattened tissue layer removed from the filter was recorded during the tests, where the functional filter is folded in zigzags and 25 times thicker than a single layer. Furthermore, microscopic images showed solids deposited on the filter fabric and fuzzy spots on the tissue could indicate possible dried aerosol spots. This could be the hint supporting the hypothesis that aerosols can be effectively filtered reducing the virus load thus also risk of super-spreading of potential infection risk to an acceptable level. Beyond this research, and with the same group, measurements were made finding out the degree of reduction in potential aerosols particles in a classroom with a continuously aerosol emitting person. On that basis from this and the other optical studies, it was concluded that the spread of COVID-19 virus can be mitigated through effective air purification systems in classrooms and students can continue learning smoothly during the ongoing pandemic.