JISE

Covid-19 pandemic detection is the key to health safety and coronavirus prevention. Due to the complex changes in CT scan treatment, it is difficult to identify the Covid-19 in the lung image. According to the latest clinical research, an automated fast framework is still required to resolve error prone problem from the pandemic assessment and Covid-19 patients screening during this critical control period. Computer aided methods can be very useful in this regard. They are suitable to estimate the infected lung boundary based on elliptical Hough transform with reduced time processing. In this paper, we propose to use a computerized approach to show that the deep neural network (DNN) is a distinctive method to classify Covid-19 pandemic. Experimental results on various lung CT scan images of different Covid-19 patients, demonstrate the effectiveness of the proposed methodology when compared to the manual scoring of pathological experts. According to the performance evaluation, we recorded more than 92% for accuracy of infection detected in ROI scoring over the truths provided by experienced radiologists. Comparative automatic studies are performed to demonstrate the suitability of the proposed technique over other advanced techniques from the literature.