Pneumonia Classification for Covid-19 Based on Machine Learning

Abstract

The year 2019 ended by giving healthcare systems worldwide a catastrophic blow, leaving hospitals flooded with patients, doctors overburdened, and distressing mortality rates due to the unforeseen spread of COVID-19. On contemplating the events of the last two years and a half, hospitals that required a shorter duration to confirm the diagnosis were able to save significantly more lives. One major issue hindering the efficiency of the diagnosis process was the primary method of diagnosis, the RT-PCR test. It is time-consuming, costly, and not readily available, especially in rural areas. Consequently, physicians started relying heavily on other diagnostic tools such as chest x-rays. A web tool was developed to help physicians upload and analyze x-ray images quickly using several machine learning models in the backend to diagnose COVID-19 or pneumonia with its types. This was attained using multiple datasets and techniques such as deep learning models and pre-trained models. The web tool includes two classifications, one differentiates normal, COVID-19, and pneumonia (PCN), and the second differentiates viral and bacterial pneumonia (BV). Six pre-trained models were used for each classification with average 95 percent testing accuracy for PCN and 85 percent for BV. Our own proposed model had a 90 percent testing accuracy for PCN and 80 percent for BV. The results presented in this paper can be used to aid in ameliorating the efficiency of healthcare systems in the diagnosis of COVID-19 and pneumonia.

Machine Learning First Response to COVID-19: A Systematic Literature Review of Clinical Decision Assistance Approaches during Pandemic Years from 2020 to 2022

“…This approach also has value as a preliminary screening tool aiming to diminish the workload on hospital staff and reduce the rate of misdiagnosis of patients with COVID-19 [178][179][180]. The enhanced prediction of disease severity via AI on CT images allows improved mortality prediction [126,129,[181][182][183][184][185][186][187] and discrimination from other forms of pneumonia not due to SARS-CoV-2 [188][189][190]. Some works [79,117] created a radiomicsand DL-based model showing the robustness of the approach on data from several sites.…”

From Pixels to Pathology: Employing Computer Vision to Decode Chest Diseases in Medical Images

“…The dataset analyzed comprised 10,293 CXR scans, including 2,874 with COVID-19, 4,200 with pneumonia, and 3,218 categorized as normal. The data was sourced from the COVID Chest repository on Kaggle and an X-ray dataset [53].…”

A diagnosis model of dementia via machine learning

“…The combination of machine learning and the medical field is particularly remarkable. Especially in the severe epidemic period, the use of machine learning methods can help doctors to quickly identify lung CT images and make a diagnosis ( Elaziz et al, 2020 ; Prakash et al, 2020 ; Afshar et al, 2021 ; Aboghazalah et al, 2022 ; Das et al, 2022 ; Elkamouny and Ghantous, 2022 ; Shiri et al, 2022 ; Sun et al, 2022 ). There are also many studies and applications of machine learning in dementia, researchers also summarize many applications of machine learning and deep learning in dementia ( Ahmed et al, 2018 ; Miah et al, 2021 ).…”