Selected article for: "convolutional neural network and deep CNNs convolutional neural network"
Author: Qiblawey, Yazan; Tahir, Anas; Chowdhury, Muhammad E. H.; Khandakar, Amith; Kiranyaz, Serkan; Rahman, Tawsifur; Ibtehaz, Nabil; Mahmud, Sakib; Al-Madeed, Somaya; Musharavati, Farayi
Title: Detection and severity classification of COVID-19 in CT images using deep learning
  • Cord-id: mk3v6vav
  • Document date: 2021_2_15
    • ID: mk3v6vav
    Snippet: Since the breakout of coronavirus disease (COVID-19), the computer-aided diagnosis has become a necessity to prevent the spread of the virus. Detecting COVID-19 at an early stage is essential to reduce the mortality risk of the patients. In this study, a cascaded system is proposed to segment the lung, detect, localize, and quantify COVID-19 infections from computed tomography (CT) images Furthermore, the system classifies the severity of COVID-19 as mild, moderate, severe, or critical based on
    Document: Since the breakout of coronavirus disease (COVID-19), the computer-aided diagnosis has become a necessity to prevent the spread of the virus. Detecting COVID-19 at an early stage is essential to reduce the mortality risk of the patients. In this study, a cascaded system is proposed to segment the lung, detect, localize, and quantify COVID-19 infections from computed tomography (CT) images Furthermore, the system classifies the severity of COVID-19 as mild, moderate, severe, or critical based on the percentage of infected lungs. An extensive set of experiments were performed using state-of-the-art deep Encoder-Decoder Convolutional Neural Networks (ED-CNNs), UNet, and Feature Pyramid Network (FPN), with different backbone (encoder) structures using the variants of DenseNet and ResNet. The conducted experiments showed the best performance for lung region segmentation with Dice Similarity Coefficient (DSC) of 97.19% and Intersection over Union (IoU) of 95.10% using U-Net model with the DenseNet 161 encoder. Furthermore, the proposed system achieved an elegant performance for COVID-19 infection segmentation with a DSC of 94.13% and IoU of 91.85% using the FPN model with the DenseNet201 encoder. The achieved performance is significantly superior to previous methods for COVID-19 lesion localization. Besides, the proposed system can reliably localize infection of various shapes and sizes, especially small infection regions, which are rarely considered in recent studies. Moreover, the proposed system achieved high COVID-19 detection performance with 99.64% sensitivity and 98.72% specificity. Finally, the system was able to discriminate between different severity levels of COVID-19 infection over a dataset of 1,110 subjects with sensitivity values of 98.3%, 71.2%, 77.8%, and 100% for mild, moderate, severe, and critical infections, respectively.

    Search related documents:
    Co phrase search for related documents
    • accurate diagnosis and lstm network: 1
    • accurate diagnosis and lstm network short term memory: 1
    • accurate diagnosis and lung background: 1, 2, 3, 4, 5
    • accurately predict and long lstm network short term memory: 1, 2
    • accurately predict and lstm network: 1
    • accurately predict and lstm network short term memory: 1, 2
    • accurately predict and lung background: 1
    • adaptive region and long lstm network short term memory: 1
    • adaptive region and lstm network: 1
    • adaptive region and lstm network short term memory: 1
    • long lstm network short term memory and lstm network: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25
    • long lstm network short term memory and lstm network short term memory: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25
    • long lstm network short term memory and lung background: 1
    • longitudinal monitoring and lung background: 1
    • lstm network and lung background: 1