2021/04/11 by Shubham Chaudhary, Chaudhary, Shubham, Vinit Jakhetiya +8
Medicine · #COVID-19 diagnosis using AI #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #FOS: Electrical engineering #Image and Video Processing (eess.IV) #Machine Learning (cs.LG) #Pneumonia and Respiratory Infections #Radiomics and Machine Learning in Medical Imaging #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2104.05121
openalex publication_date 2021/04/11 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
We propose a two-stage Convolutional Neural Network (CNN) based\nclassification framework for detecting COVID-19 and Community-Acquired\nPneumonia (CAP) using the chest Computed Tomography (CT) scan images. In the\nfirst stage, an infection - COVID-19 or CAP, is detected using a pre-trained\nDenseNet architecture. Then, in the second stage, a fine-grained three-way\nclassification is done using EfficientNet architecture. The proposed\nCOVID+CAP-CNN framework achieved a slice-level classification accuracy of over\n94% at identifying COVID-19 and CAP. Further, the proposed framework has the\npotential to be an initial screening tool for differential diagnosis of\nCOVID-19 and CAP, achieving a validation accuracy of over 89.3% at the finer\nthree-way COVID-19, CAP, and healthy classification. Within the IEEE ICASSP\n2021 Signal Processing Grand Challenge (SPGC) on COVID-19 Diagnosis, our\nproposed two-stage classification framework achieved an overall accuracy of 90%\nand sensitivity of .857, .9, and .942 at distinguishing COVID-19, CAP, and\nnormal individuals respectively, to rank first in the evaluation. Code and\nmodel weights are available at\nhttps://github.com/shubhamchaudhary2015/ctcovid19capcnn\n