Diagnosis of COVID-19 from X-rays Using Combined CNN-RNN Architecture with Transfer Learning

Abstract

The confrontation of COVID-19 pandemic has become one of the promising challenges of the world healthcare. Accurate and fast diagnosis of COVID-19 cases is essential for correct medical treatment to control this pandemic. Compared with the reverse-transcription polymerase chain reaction (RT-PCR) method, chest radiography imaging techniques are shown to be more effective to detect coronavirus. For the limitation of available medical images, transfer learning is better suited to classify patterns in medical images. This paper presents a combined architecture of convolutional neural network (CNN) and recurrent neural network (RNN) to diagnose COVID-19 from chest X-rays. The deep transfer techniques used in this experiment are VGG19, DenseNet121, InceptionV3, and Inception-ResNetV2. CNN is used to extract complex features from samples and classified them using RNN. The VGG19-RNN architecture achieved the best performance among all the networks in terms of accuracy and computational time in our experiments. Finally, Gradient-weighted Class Activation Mapping (Grad-CAM) was used to visualize class-specific regions of images that are responsible to make decision. The system achieved promising results compared to other existing systems and might be validated in the future when more samples would be available. The experiment demonstrated a good alternative method to diagnose COVID-19 for medical staff.

Detection of COVID-19 from CT Lung Scans Using Transfer Learning

“…In a paper by Islam et al [17], an ensemble CNN-RNN architecture is proposed that investigates and compares the use of a VGG-19, DenseNet121, InceptionV3, and Incep-tionResNetV2 convolutional base, used for segmentation and feature extraction in combination with a recurrent neural network (RNN) classifier. e study aimed to create a framework that is capable of distinguishing between COVID-19 lung scans, community acquired pneumonia lung scans, and healthy lung scans.…”

“…Accuracy % Recall % MobileNet-v2 [26] 97.40 99.10 VGG-19 [26] 98.75 92.85 VGG-16 [11] 95.88 96.00 EfficientNet-B4 [12] 96.70 96.69 DenseNet-201 [10] 97.00 -VGG-19 + RNN [17] 99.9 99.8 XCeption [27] 99.52 99.12 ResNet-101 [29] 99.51 100 VGG-19 + CLAHE (proposed) 95.75 97. 13 Computational Intelligence and Neuroscience…”

“…e only metric reported in this study is the accuracy of the trained model. Computational Intelligence and Neuroscience Islam et al [17] investigated the use of CNN-RNN ensemble architectures in this task. e use of the VGG-19 architecture produces the highest accuracies and recalls.…”

Wearable Technology to Assist the Patients Infected with Novel Coronavirus (COVID-19)

“…As the vaccine for the novel coronavirus is still under trial, efforts should be focused on preventing the spread of the virus. In the present era of artificial intelligence (AI) [17][18][19][20], the new technologies can play a vital role to assist the world against the pandemic. Wearable technology that can collect a wide range of data such as heart rate, blood pressure, body temperature, ECG, lung sound, blood oxygen saturation (SpO 2 ) level, etc.…”

Deep Learning Applications to Combat Novel Coronavirus (COVID-19) Pandemic

“…From the time of civilization, several diseases like heart disease [5], diabetes [6], liver disorder [7], breast cancer [8][9][10], COVID-19 [11][12][13], etc. caused severe and acute actions on human health, and artificial intelligencebased systems show better performance to identify those diseases.…”