Early screening of COVID-19 from chest CT using deep learning technique

ABSTRACT

Coronaviruses are mainly a big family of viruses that are highly capable of causing illness both in animals and humans. The scientific name of the most recently discovered corona virus disease is COVID-19. Most of the countries are performing the manual testing which is beneficial to know the actual situation, feature of the disease, so that appropriate decision can be taken. The main drawbacks of manual testing is that it is very expensive, sparse availability of testing kits, inefficient blood test, and minimum 5–6 h will require to generate the report of blood test. So in these circumstances, deep learning plays a crucial role to detect the positive cases as early as possible so as to prevent the further spread of this epidemic and provide fast and efficient treatment to the effected patients. The developed model consists of three groups COVID-19, Influenza-A viral pneumonia, and healthy cases. Our proposed detection model got 98.78% accuracy. In this study, we propose a fast and efficient way to identify COVID-19 patients with multi-task deep learning (DL) methods from CT scan images. We have developed two models (a) Inception residual recurrent convolutional neural network with transfer learning (TL) approach for COVID-19 detection and (b) NABLA-N network model for segmenting the regions infected by COVID-19. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2021.