Detection of COVID-19 Using Transfer Learning and Grad-CAM Visualization on Indigenously Collected X-ray Dataset

Abstract

The COVID-19 outbreak began in December 2019 and has dreadfully affected our lives since then. More than three million lives have been engulfed by this newest member of the corona virus family. With the emergence of continuously mutating variants of this virus, it is still indispensable to successfully diagnose the virus at early stages. Although the primary technique for the diagnosis is the PCR test, the non-contact methods utilizing the chest radiographs and CT scans are always preferred. Artificial intelligence, in this regard, plays an essential role in the early and accurate detection of COVID-19 using pulmonary images. In this research, a transfer learning technique with fine tuning was utilized for the detection and classification of COVID-19. Four pre-trained models i.e., VGG16, DenseNet-121, ResNet-50, and MobileNet were used. The aforementioned deep neural networks were trained using the dataset (available on Kaggle) of 7232 (COVID-19 and normal) chest X-ray images. An indigenous dataset of 450 chest X-ray images of Pakistani patients was collected and used for testing and prediction purposes. Various important parameters, e.g., recall, specificity, F1-score, precision, loss graphs, and confusion matrices were calculated to validate the accuracy of the models. The achieved accuracies of VGG16, ResNet-50, DenseNet-121, and MobileNet are 83.27%, 92.48%, 96.49%, and 96.48%, respectively. In order to display feature maps that depict the decomposition process of an input image into various filters, a visualization of the intermediate activations is performed. Finally, the Grad-CAM technique was applied to create class-specific heatmap images in order to highlight the features extracted in the X-ray images. Various optimizers were used for error minimization purposes. DenseNet-121 outperformed the other three models in terms of both accuracy and prediction.

Data-Driven Analytics Leveraging Artificial Intelligence in the Era of COVID-19: An Insightful Review of Recent Developments

“…Pinter et al [106] Multi-layered perceptron Predictions of mortality rate and infected cases Aminu et al [107] Deep neural networks Detection of people with COVID-19 Magar et al [108] Ensemble techniques Virus-antibody sequence analysis and patients' Identification Zeng et al [109] Extreme Gradient Boosting (XGBoost) Forecasting of patient survival probability Ashraf et al [110] Machine & deep learning models Predict the severity of disease or chances of death Shah et al [111] Convolutional neural network (CNN) COVID-19 detection from X-ray images Prakash et al [112] Autoregressive Integrated Moving Average Impact analysis of various policies Rathod et al [113] AI Prediction models Effective crisis preparedness and management Ullah et al [114] Logistic Regression and Support Vector Machine Classification of patients with/without COVID-19 Rathod et al [115] SVM, RProp, and Decision tree Detection of abnormal data for effective analysis Hu et al [116] Spectral Clustering (SC) algorithm Feasible analysis model for the treatment & diagnosis Rashed et al [117] Long short-term memory (LSTM) network Provides public awareness about the risks of COVID-19 Singh et al [118] ResNet152V2 and VGG16 CNN Reduce the high false-negative results of the RT-PCR Saverino et al [119] Digital and artificial intelligence platform (DAIP) Changes implementation in rehabilitation services Peddinti et al [120] Convolutional Neural Network (CNN) Detection of COVID-19 cases in public places Malla et al [121] Ensemble deep learning model Real-time sentiment analysis of COVID-19 data Lella et al [122] Convolutional Neural Network (CNN) model Respiratory sound classification for patient identification Haleem et al [123] Artificial neuronal networks (ANN) Predictions of survival of COVID-19 patients Hashimi et al [124] Deep learning models Tracking and identifying potential virus spreaders Amaral et al [125] Artificial neuronal networks (ANN) forecasting and monitoring the progress of Covid-19 Zgheib et al [126] Collection of ensemble learning methods Detecting COVID-19 virus based on patient's demographics Ferrari et al [127] Bayesian framework Predictions about the behavior of the COVID-19 epidemic Almalki et al [128] COVID Inception-ResNet model (CoVIRNet) Automatic diagnosis of the COVID-19 patients Umair et al …”

Neonatal Jaundice Diagnosis Using a Smartphone Camera Based on Eye, Skin, and Fused Features with Transfer Learning

“…In fact, artificial intelligence, specifically machine learning and deep learning, has contributed to tackling multiple challenges in the diagnosis of different diseases. Since their emergence, a wide range of research has been carried out with breakthrough results [2][3][4][5]. When symptoms are visible, images of the infected area are collected, and computer vision and image processing algorithms and techniques are applied to extract features that are fed to the diagnostic models.…”

“…Instead, they check both areas simultaneously for further assurance. Moreover, previous studies in this domain have not applied deep learning and transfer learning models, despite their success in the diagnosis of different diseases [2][3][4][5]. In this work, we seek to fill this gap.…”

Noninvasive Detection of Respiratory Disorder Due to COVID-19 at the Early Stages in Saudi Arabia

“…Hence, the proposed method can help mitigate the ongoing spread of COVID-19 and can continuously and seamlessly monitor the specific patients. In the literature [11][12][13][14][15][16], COVID-19 monitoring/prediction methods are available; however, these methods [11][12][13][14][15][16][17] are mainly focused on classification without any encryption. The method outlined in [11][12][13][14][15][16] mainly uses AIand ML-based techniques.…”

“…In the literature [11][12][13][14][15][16], COVID-19 monitoring/prediction methods are available; however, these methods [11][12][13][14][15][16][17] are mainly focused on classification without any encryption. The method outlined in [11][12][13][14][15][16] mainly uses AIand ML-based techniques. A number of ML techniques are used for other applications, and details can be found in [18,19].…”