Abstract
Since December 2019, the pandemic COVID-19 has been connected to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Early identification and diagnosis are essential goals for health practitioners because early symptoms correlate with those of other common illnesses including the common cold and flu. RT-PCR is frequently used to identify SARS-CoV-2 viral infection. Although this procedure can take up to 2 days to complete and sequential monitoring may be essential to figure out the potential of false-negative findings, RT-PCR test kits are apparently in low availability, highlighting the urgent need for more efficient methods of diagnosing COVID-19 patients. Artificial intelligence (AI)-based healthcare models are more effective at diagnosing and controlling large groups of people. Hence, this paper proposes a novel AI-enabled SARS detection framework. Here, the input CT images are collected and preprocessed using a block-matching filter and histogram equalization (HE). Segmentation is performed using Compact Entropy Rate Superpixel (CERS) technique. Features of segmented output are extracted using Histogram of Gradient (HOG). Feature selection is done using Principal Component Analysis (PCA). The suggested Random Sigmoidal Artificial Neural Networks (RS-ANN) based classification approach effectively diagnoses the existence of the disease. The performance of the suggested Artificial intelligence model is analyzed and related to existing approaches. The suggested AI system may help identify COVID-19 patients more quickly than conventional approaches.
Keywords: Principal Component Analysis; Random Sigmoidal Artificial Neural Networks; artificial intelligence; block-matching filter; compact entropy rate superpixel; histogram equalization; histogram of gradient; severe acute respiratory syndrome coronavirus 2.
【저자키워드】 artificial intelligence, principal component analysis, Severe acute respiratory syndrome coronavirus 2., Random Sigmoidal Artificial Neural Networks, block-matching filter, compact entropy rate superpixel, histogram equalization, histogram of gradient, 【초록키워드】 COVID-19, SARS-CoV-2, viral infection, coronavirus, pandemic, artificial intelligence, Diagnosis, Symptom, severe acute respiratory syndrome Coronavirus, RT-PCR, principal component analysis, Segmentation, Health, artificial, healthcare, network, common cold, flu, group, false-negative, COVID-19 patients, diagnose, Feature selection, RT-PCR test, Entropy, PCA, COVID-19 patient, Principal component, acute respiratory syndrome, acute respiratory syndrome coronavirus, acute respiratory syndrome coronavirus 2, approaches, help, component, diagnosing, histogram, SARS-CoV-2 viral infection, Practitioner, Complete, approach, principal, effective, Gradient, analyzed, identify, performed, collected, Rate, the disease, suggested, illness, Neural, highlighting, Compact, 【제목키워드】 artificial, network, crisis, Neural,