Early-stage COVID-19 diagnosis in presence of limited posteroanterior chest X-ray images via novel Pinball-OCSVM Sanjay Kumar Sonbhadra a,∗ , Sonali Agarwal a , P. Nagabhushan a a Indian Institute of Information Technology, Allahabad, India, 211015 Abstract The respiratory coronavirus disease 2019 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that has been declared as pandemic by the world health organization (WHO) in March, 2020. It is evident that the infection with this coronavirus starts with the upper respiratory tract and as the virus grows, the infection can progress to lungs and develop pneumonia. According to the statistics, approximately 14% of the infected people with COVID-19 have severe cough and shortness of breath due to pneumonia, because as the viral infection increases, it damages the alveoli (small air sacs) and surrounding tissues. When the alveoli are damaged, there is an influx of liquid which is mostly inflamed cells and protein that leads to pneumonia. The conventional way of COVID- 19 diagnosis is reverse transcription polymerase chain reaction (RT-PCR), which is less sensitive during early stages specially, if the patient is asymptomatic that may further lead to more severe pneumonia. To overcome this problem an early diagnosis method is proposed in this paper via one-class classification approach using a novel pinball loss function based one-class support vector machine (PB-OCSVM) considering posteroanterior chest X-ray images. Recently, several automated COVID-19 diagnosis models have been proposed based on various deep learning architectures to identify pulmonary infections using publicly available chest X-ray (CXR) and computed tomography (CT) of corona positive patients for early diagnosis, better treatment and quick cure. In these datasets, presence of less number of COVID-19 positive samples compared to other classes (normal, pneumonia and Tuberculosis) raises the challenge for unbiased learning in deep learning models that has been solved using class balancing techniques which however should be avoided in any medical diagnosis process. Inspired by this phenomenon, this article proposes a novel PB-OCSVM model to work in presence of limited COVID-19 positive CXR samples with objectives to maximize the learning efficiency while minimize the false-positive and false-negative predictions. The proposed model outperformed over recently published deep learning approaches where accuracy, precision, specificity, sensitivity and confidence interval are used as performance measure parameters. Keywords: COVID-19, Chest X-ray, Pneumonia, Classification, Deep learning, Pinball loss, One-class support vector machine. 1. Introduction The SARS-CoV-2 is the most recently identified mem- ber of the coronavirus family causing COVID-19 [1]. This respiratory disease started for Wuhan, China during late December, 2020 and spread to all the countries worldwide [2]. World health organization (WHO) declared this infec- tious disease as a public health emergency of international concern (PHEIC) on January 30, 2020 as it reached to more than 18 countries [3] and on Feb 11, 2020 named this “COVID-19” and later, declared this a pandemic on March 11, 2020 [4, 5]. After the ten months journey, this virus caused over 37 million positive cases and more than 1 million deaths worldwide till October 10, 2020 [6]. This disease is highly infectious and to control its spread, fol- lowing three ways have been suggested as most desirable * Corresponding author: Tel.: +91 9827873773; Email address: rsi2017502@iiita.ac.in (Sanjay Kumar Sonbhadra ) preventive measures: social distancing [7], use of mask [8] and early identification of infected people. Till now, there is no best-known medication or vac- cine is available for the proper treatment of COVID-19; therefore to stop the infection the infected person must be quarantined to break the chain of infection. Real-time re- verse transcription-polymerase chain reaction (RT-PCR) is the best known method of COVID-19 diagnosis. This came into effect after the rapid antibody tests showed un- reliable results, because antibodies appear after 9-28 days of the infection and by this time, an infected person may spread the disease, if not isolated [9]. The RT-PCR test is a costly and time taking process, whereas inability to early-stage diagnosis and due to rapid growth of infec- tions, medical experts are continuously trying to search some other means of diagnostic measures. SARS-CoV-2 virus initially affects the respiratory sys- tem of the infected person and in later stages, it affects lungs that may cause severe pneumonia [10]. It is evident Preprint submitted to Elsevier November 12, 2021 arXiv:2010.08115v1 [eess.IV] 16 Oct 2020