Cascaded deep learning classifiers for computer-aided diagnosis of COVID-19 and pneumonia diseases in X-ray scans

Abstract

Computer-aided diagnosis (CAD) systems are considered a powerful tool for physicians to support identification of the novel Coronavirus Disease 2019 (COVID-19) using medical imaging modalities. Therefore, this article proposes a new framework of cascaded deep learning classifiers to enhance the performance of these CAD systems for highly suspected COVID-19 and pneumonia diseases in X-ray images. Our proposed deep learning framework constitutes two major advancements as follows. First, complicated multi-label classification of X-ray images have been simplified using a series of binary classifiers for each tested case of the health status. That mimics the clinical situation to diagnose potential diseases for a patient. Second, the cascaded architecture of COVID-19 and pneumonia classifiers is flexible to use different fine-tuned deep learning models simultaneously, achieving the best performance of confirming infected cases. This study includes eleven pre-trained convolutional neural network models, such as Visual Geometry Group Network (VGG) and Residual Neural Network (ResNet). They have been successfully tested and evaluated on public X-ray image dataset for normal and three diseased cases. The results of proposed cascaded classifiers showed that VGG16, ResNet50V2, and Dense Neural Network (DenseNet169) models achieved the best detection accuracy of COVID-19, viral (Non-COVID-19) pneumonia, and bacterial pneumonia images, respectively. Furthermore, the performance of our cascaded deep learning classifiers is superior to other multi-label classification methods of COVID-19 and pneumonia diseases in previous studies. Therefore, the proposed deep learning framework presents a good option to be applied in the clinical routine to assist the diagnostic procedures of COVID-19 infection.

Convolutional Neural Network Model to Detect COVID-19 Patients Utilizing Chest X-ray Images

“…Most of the works were focused on high sensitivity, i.e. how the classifiers can identify COVID-19 positive cases frequently (Karar et al, 2020; Karthik et al, 2020; Khan et al, 2020; Ismael & Şengür, 2021; Ohata et al, 2021). But, nowadays the community transmission is a great issue to prevent the spread of COVID-19 and the growth of false negative rates are accelerated, which is a great concern.…”

“…In this work, we verified experimental results using various evaluation metrics like accuracy, AUC, F-measure, sensitivity, and specificity respectively. Several works have analyzed a few number of COVID-19 samples along with other cases where the experimental dataset was remained imbalanced (Karar et al, 2020; Sekeroglu & Ozsahin, 2020; Shankar & Perumal, 2020; Zebin & Rezvy, 2020). Sometimes, they were conducted with a separate dataset where the scarcity of samples was found in both of these datasets (Apostolopoulos & Mpesiana, 2020; Shankar & Perumal, 2020).…”

“…Apostolopoulos & Mpesiana (2020) implemented transfer learning using CNNs into a small medical image dataset (1427 X-ray images), which provided highest 96.78% accuracy, 98.66% sensitivity, and 96.46% specificity respectively. Karar et al (2020) proposed a deep CNN architecture called COVIDX-Net that investigated 50 chest X-ray images with 25 COVID-19 cases and provided 90% accuracy and 91% F-score. Minaee et al (2020) proposed a deep learning framework based on 5000 images named COVID-Xray-5k where they applied ResNet18, ResNet50, SqueezeNet and Densenet-121 into them and produced sensitivity 97.5% and specificity 90% on average.…”

Secure CT-Image Encryption for COVID-19 Infections Using HBBS-Based Multiple Key-Streams

“…According to the WHO scientific guidelines for national laboratories, the Real-Time Polymerase Chain Reaction (RT-PCR) assay is the gold standard for verifying the diagnosis of COVID-19 [ 5 ]. However, the RT-PCR test becomes insufficient, and showed high levels results of false-negative to confirm positive COVID-19 cases.…”

A medical decision support system for predicting the severity level of COVID-19

“…Its response against COVID-19 includes three phases: (1) big data analytics: by gathering national health insurance database and integrating with immigration and customs database; (2) implementing new technology: using QR code scanning and online reporting to classify travelers’ infectious risks based on-flight origin and travel history in the past 14 days; (3) proactive testing: to amplify the COVID-19 case finding [ 2 , 3 ]. Karar et al [ 14 ] using X-ray scans proposed computer-aided diagnosis (CAD) systems based on cascaded deep learning (DL) classifiers for COVID-19. A similar study, analyzing chest X-ray images, has been conducted by Shankar and Perumal [ 15 ] for COVID-19 diagnosis and classification using a novel hand-crafted with DL feature-based fusion model.…”