COVID-19 Diagnosis from Chest X-ray Images Using a Robust Multi-Resolution Analysis Siamese Neural Network with Super-Resolution Convolutional Neural Network

Abstract

Chest X-ray (CXR) is becoming a useful method in the evaluation of coronavirus disease 19 (COVID-19). Despite the global spread of COVID-19, utilizing a computer-aided diagnosis approach for COVID-19 classification based on CXR images could significantly reduce the clinician burden. There is no doubt that low resolution, noise and irrelevant annotations in chest X-ray images are a major constraint to the performance of AI-based COVID-19 diagnosis. While a few studies have made huge progress, they underestimate these bottlenecks. In this study, we propose a super-resolution-based Siamese wavelet multi-resolution convolutional neural network called COVID-SRWCNN for COVID-19 classification using chest X-ray images. Concretely, we first reconstruct high-resolution (HR) counterparts from low-resolution (LR) CXR images in order to enhance the quality of the dataset for improved performance of our model by proposing a novel enhanced fast super-resolution convolutional neural network (EFSRCNN) to capture texture details in each given chest X-ray image. Exploiting a mutual learning approach, the HR images are passed to the proposed Siamese wavelet multi-resolution convolutional neural network to learn the high-level features for COVID-19 classification. We validate the proposed COVID-SRWCNN model on public-source datasets, achieving accuracy of 98.98%. Our screening technique achieves 98.96% AUC, 99.78% sensitivity, 98.53% precision, and 98.86% specificity. Owing to the fact that COVID-19 chest X-ray datasets are low in quality, experimental results show that our proposed algorithm obtains up-to-date performance that is useful for COVID-19 screening.

Keywords: COVID-19; Siamese network; chest X-ray (CXR); convolutional neural network; multi-resolution analysis; super resolution.

Figure 1

Data collection of chest X-ray images of different pneumonia-related illnesses including COVID-19.

Figure 2

This figure shows the network architecture of the SRCNN, FSRCNN, and our proposed enhanced super-resolution framework called EFSRCNN. In a logical sense, our proposed model is centered on the merits of both the SRCNN and FSRCNN. First, EFSRCNN uses the bicubic interpolated version of the ground-truth low-resolution image as an input, similar to the process in SRCNN but different from the process in FSRCNN. Similar to FSRCNN, a deconvolutional layer is added at the end of the network to achieve up-sampling. Shrinking, mapping, and dilation phases of EFSRCNN replace the non-linear mapping phase in SRCNN and it is quite similar to the phases in FSRCNN. Nevertheless, EFSRCNN has a deeper network topology compared to FSRCNN. The sizes of the filters within the mapping layers are kept similar to FSRCNN. These enhancements give EFSRCNN higher performance while lowering the computational cost compared to SRCNN and FSRCNN.

Figure 3

Our proposed Siamese wavelet multi-resolution convolutional neural network.

Figure 4

The proposed super-resolution-based Siamese wavelet multi-resolution convolutional neural network for COVID-19 classification (COVID-SRWCNN).

Figure 5

Comparison of the quantitative results of our proposed EFSRCNN with other selected state-of-the-art models using the same dataset. The PSNR value is reported on the left while the SSIM value is reported on the right for the whole region.

Figure 6

Comparison of the quantitative results of our proposed EFSRCNN with other selected state-of-the-art models using the same dataset. The PSNR value is reported on the left while the SSIM value is reported on the right for the region of interest.

Figure 7

Performance report of our model and selected pre-trained models. (a) Sensitivity report for the selected deep pre-trained models and our proposed model. (b) Specificity report for the selected deep pre-trained models and our proposed model.

Figure 8

Training and validation report of our model with and without super resolution (SR). (a) Accuracy curves showing the performance of our proposed COVID-SRWCNN with and without super resolution (SR). (b) Loss curves reported for our proposed COVID-SRWCNN with and without super resolution (SR).

Figure 9

Test report of our model with and without super resolution (SR). (a) Test accuracy curves showing the performance of our proposed COVID-SRWCNN with and without super resolution (SR). (b) Test loss curves reported for our proposed COVID-SRWCNN with and without super resolution (SR).

Figure 10

Comparison report for the selected state-of-the-art COVID-19 models and our proposed model. (a) Accuracy report for the selected state-of-the-art COVID-19 models and our proposed model. (b) Sensitivity report for the selected state-of-the-art COVID-19 models and our proposed model.

Figure 11

Comparison report for the selected deep pre-trained models and our proposed model. (a) Accuracy report for the selected deep pre-trained models and our proposed model. (b) AUC report for the selected deep pre-trained models and our proposed model.

Figure 12

Comparison report for the selected state-of-the-art COVID-19 models and our proposed model. (a) Specificity report for the selected state-of-the-art COVID-19 models and our proposed model. (b) AUC report for the selected state-of-the-art COVID-19 models and our proposed model.

Figure 13

Comparison report of our proposed COVID-SRWCNN with and without super resolution (SR). (a) ROC–AUC curves of our proposed COVID-SRWCNN with and without super resolution (SR). (b) Precision–recall curves of our proposed COVID-SRWCNN with and without super resolution (SR).

Figure 14

Comparison report of our proposed COVID-SRWCNN in comparison with selected state-of-the-art COVID-19 models using the same dataset. (a) ROC–AUC curves of our proposed COVID-SRWCNN in comparison with selected state-of-the-art COVID-19 models using the same dataset. (b) Precision–recall curves of our proposed COVID-SRWCNN in comparison with selected state-of-the-art COVID-19 models using the same dataset.