A Novel Deep Learning Framework for Pulmonary Embolism Detection for Covid-19 Management

Pulmonary Embolism is a blood clot in the lung which restricts flow and reduces oxygen level resulting mortality if it untreated. Further, pulmonary embolism evidenced prominently segmental sub-segmental regions of computed tomography angiography images COVID-19 patients. detection from these significant research problem challenging pandemic venture early disease detection, treatment, prognosis. Inspired by several investigations based on deep learning this context, two-stage framework has been proposed for realized as segmentation model. It implemented cascade convolutional superpixel neural network regularized UNet candidates well embolisms, respectively. The model tested with two public datasets achieved testing accuracy 99%. demonstrates high sensitivities 88.43%, 88.36%, 89.93% at 0, 2, 5 mm localization errors, respectively false positives they are superior to state-of-the-art models, signifying potential applications treatment protocols diverse diseases COVID-19.