COVIDX-LwNet: A Lightweight Network Ensemble Model for the Detection of COVID-19 Based on Chest X-ray Images.

Recently, the COVID-19 pandemic coronavirus has put a lot of pressure on health systems around the world. One of the most common ways to detect COVID-19 is to use chest X-ray images, which have the advantage of being cheap and fast. However, in the early days of the COVID-19 outbreak, most studies applied pretrained convolutional neural network (CNN) models, and the features produced by the last convolutional layer were directly passed into the classification head. In this study, the proposed ensemble model consists of three lightweight networks, Xception, MobileNetV2 and NasNetMobile as three original feature extractors, and then three base classifiers are obtained by adding the coordinated attention module, LSTM and a new classification head to the original feature extractors. The classification results from the three base classifiers are then fused by a confidence fusion method. Three publicly available chest X-ray datasets for COVID-19 testing were considered, with ternary (COVID-19, normal and other pneumonia) and quaternary (COVID-19, normal) analyses performed on the first two datasets, bacterial pneumonia and viral pneumonia classification, and achieved high accuracy rates of 95.56% and 91.20%, respectively. The third dataset was used to compare the performance of the model compared to other models and the generalization ability on different datasets. We performed a thorough ablation study on the first dataset to understand the impact of each proposed component. Finally, we also performed visualizations. These saliency maps not only explain key prediction decisions of the model, but also help radiologists locate areas of infection. Through extensive experiments, it was finally found that the results obtained by the proposed method are comparable to the state-of-the-art methods.

IoT Deployable Lightweight Deep Learning Application For COVID-19 Detection With Lung Diseases Using RaspberryPi

Proper assessment of COVID-19 patients has become critical to mitigating and halting the disease's rapid expansion during the present COVID-19 epidemic across the nations. Due to the presence of chronic lung/pulmonary diseases, the intensity and demise rates of COVID-19 patients were increased. This study will analyze radiography utilizing chest X-ray images (CXI), one of the most successful testing methods for COVID-19 case identification. Given that deep learning (DL) is a useful method and technique for image processing, there have been several research on COVID-19 case identification using CXI to train DL models. While few of the study claims outstanding predictive outcomes, their suggested models may struggle with overfitting, excessive variance, and generalization mistakes due to noise, a limited number of datasets and could not be deployed to IoT devices due to heavy network size. Considering deep Convolutional Neural Network (CNN) can conquer the weaknesses by getting predictions with several diseases using a single model deployed on a real-time IoT device. We propose a lightweight Deep Learning model (LDC-Net) that has spearheaded an open-sourced COVID-19 case identification technique using CNN-generated CXI by utilizing a suggested strategy aware of distinct features learning of different classes. Experimental results on Raspberry Pi show that LDC-Net provides encouraging outputs for detecting COVID-19 cases with an overall 96.86% precision, 96.78% recall, 96.77% F1-score, and 99.28% accuracy, better than other state-of-the-art models. By empowering the Internet of Things-IoT and IoMT devices, this suggested framework can identify COVID-19 from CXI and other seven lung diseases with healthy labels. © 2022 IEEE.