COV-DLS: Prediction of COVID-19 from X-Rays Using Enhanced Deep Transfer Learning Techniques

Abstract

In this paper, modifications in neoteric architectures such as VGG16, VGG19, ResNet50, and InceptionV3 are proposed for the classification of COVID-19 using chest X-rays. The proposed architectures termed “COV-DLS” consist of two phases: heading model construction and classification. The heading model construction phase utilizes four modified deep learning architectures, namely Modified-VGG16, Modified-VGG19, Modified-ResNet50, and Modified-InceptionV3. An attempt is made to modify these neoteric architectures by incorporating the average pooling and dense layers. The dropout layer is also added to prevent the overfitting problem. Two dense layers with different activation functions are also added. Thereafter, the output of these modified models is applied during the classification phase, when COV-DLS are applied on a COVID-19 chest X-ray image data set. Classification accuracy of 98.61% is achieved by Modified-VGG16, 97.22% by Modified-VGG19, 95.13% by Modified-ResNet50, and 99.31% by Modified-InceptionV3. COV-DLS outperforms existing deep learning models in terms of accuracy and F1-score.

IFCnCov: An IoT‐based smart diagnostic architecture for COVID‐19

“…Transfer learning (TL), a prominent DL training approach, adjusts a learned model to a new goal. 10 Residual networks, or ResNets, are a type of ANN and TL that skip-connects hidden nodes to create a more complex NN with no vanishing gradients in its upper layers. Its type includes ResNet-18, ResNet23, ResNet-34, ResNet-50, ResNet101 and so forth where layer numbers indicate the same architecture and this research uses 101-layered ResNet101.…”

“…Transfer learning (TL), a prominent DL training approach, adjusts a learned model to a new goal 10 . Residual networks, or ResNets, are a type of ANN and TL that skip‐connects hidden nodes to create a more complex NN with no vanishing gradients in its upper layers.…”

NSCGCN: A novel deep GCN model to diagnosis COVID-19

“…We consider the limited number of labelled training samples insufficient to retrain the entire model from scratch. To attain high classification accuracy, we follow the transfer learning method [ 4 ] and propose a feature extraction module based on the deep CNN models trained on the ImageNet to extract the pixel-level features from medical images more effectively.…”

“…As supplementary diagnosis methods, CT and CXR can improve the diagnosis rate of COVID-19 and reduce the false-positive rate. Because of the low-cost and low-radiation dose, the CXR images have been widely used in COVID-19 diagnosis [ 4 ].…”

An Effective Deep Learning Model to Discriminate Coronavirus Disease From Typical Pneumonia

“…medical treatment processes and the procedures of preventive containment (Abdel-Basset et al, 2021). Though numerous criteria provide successful COVID-19 diagnosis for people, the tools of clinical laboratory based on RT-PCR and sequencing of virus nucleic acid suffer from various deficiencies, such as the test results become prepared at the earliest twentyfour hours concerning crucial cases and usually require many days to dedicate a decision (Kumar et al, 2022). Recently, the world health organization announced that RT-PCR may provide incorrect results in COVID-19 cases owing to low-quality specimens obtained from patients, unsuitable specimens processing, and taking specimens at the late or early stages of the disease.…”