Real-Time Multi-Class Infection Classification for Respiratory Diseases

Real-time disease prediction has emerged as the main focus of study in the field of computerized medicine. Intelligent disease identification framework can assist medical practitioners in diagnosing disease in a way that is reliable, consistent, and timely, successfully lowering mortality rates, particularly during endemics and pandemics. To prevent this pandemic's rapid and widespread, it is vital to quickly identify, confine, and treat affected individuals. The need for auxiliary computer-aided diagnostic (CAD) systems has grown. Numerous recent studies have indicated that radiological pictures contained critical information regarding the COVID-19 virus. Utilizing advanced convolutional neural network (CNN) architectures in conjunction with radiological imaging makes it possible to provide rapid, accurate, and extremely useful susceptible classifications. This research work proposes a methodology for real-time detection of COVID-19 infections caused by the Corona Virus. The purpose of this study is to offer a two-way COVID-19 (2WCD) diagnosis prediction deep learning system that is built on Transfer Learning Methodologies (TLM) and features customized fine-tuning on top of fully connected layered pre-trained CNN architectures. 2WCD has applied modifications to pre-trained models for better performance. It is designed and implemented to improve the generalization ability of the classifier for binary and multi-class models. Along with the ability to differentiate COVID19 and No-Patient in the binary class model and COVID-19, No-Patient, and Pneumonia in the multi-class model, our framework is augmented with a critical add-on for visually demonstrating infection in any tested radiological image by highlighting the affected region in the patient's lung in a recognizable color pattern. The proposed system is shown to be extremely robust and reliable for real-time COVID-19 diagnostic prediction. It can also be used to forecast other lung-related disorders. As the system can assist medical practitioners in diagnosing the greatest number of patients in the shortest amount of time, radiologists can also be used or published online to assist any less-experienced individual in obtaining an accurate immediate screening for their radiological images.

A Medical Decision Support System to Detect Covid-19 Pneumonia Using CNN

Due to the pandemic by the spread of the COVID virus, there has been a mandatory demand to screen patients. Predominantly RTPCR test is used to detect the virus. The RTPCR test is the most commonly used technique to detect COVID - 19 viruses. The test takes a minimum of 12 hours which is time-consuming and might put a patient's life at stake. This detection method for COVID screening is said to have a false detection rate. CT scans have been used for COVID-19 screening and using CT has several challenges especially since their radiation dose is considerably higher than x-rays. Hence, CXRs are a better choice for the initial assessment. Detection of COVID-19 pneumonia is a fine-grained problem as doctors cannot detect it just by looking at the x-ray images. Moreover, the radiologists visit many patients every day and the diagnosis process take significant time, which may increase errors in screening notably. Therefore, a medical decision support system for screening COVID-19 patients is of utmost importance. Our proposed system is a web application that helps to screen COVID-19 patients effectively. © 2021 IEEE.

Automatic detection of COVID-19 in chest radiographs using serially concatenated deep and handcrafted features.

Since the infectious disease occurrence rate in the human community is gradually rising due to varied reasons, appropriate diagnosis and treatments are essential to control its spread. The recently discovered COVID-19 is one of the contagious diseases, which infected numerous people globally. This contagious disease is arrested by several diagnoses and handling actions. Medical image-supported diagnosis of COVID-19 infection is an approved clinical practice. This research aims to develop a new Deep Learning Method (DLM) to detect the COVID-19 infection using the chest X-ray. The proposed work implemented two methods namely, detection of COVID-19 infection using (i) a Firefly Algorithm (FA) optimized deep-features and (ii) the combined deep and machine features optimized with FA. In this work, a 5-fold cross-validation method is engaged to train and test detection methods. The performance of this system is analyzed individually resulting in the confirmation that the deep feature-based technique helps to achieve a detection accuracy of > 92% with SVM-RBF classifier and combining deep and machine features achieves > 96% accuracy with Fine KNN classifier. In the future, this technique may have potential to play a vital role in testing and validating the X-ray images collected from patients suffering from the infection diseases.

Super-Resolution Optical Microscopy to Detect Viruses (SARS-CoV-2) in Real Time

This article describes the use of evanescent light fields to directly observe and detect the newly discovered coronavirus SARS-CoV-2 that causes COVID-19. The proposed technique provides a low-cost, fast, and highly accurate method of detection. This approach builds from previous work from the authors that enables microscopic observations of nano-objects with the accuracy of nanometers and sensitivities of the order of fraction of a nanometer. © 2022, The Society for Experimental Mechanics, Inc.

Optimized chest X-ray image semantic segmentation networks for COVID-19 early detection.

BACKGROUND: Although detection of COVID-19 from chest X-ray radiography (CXR) images is faster than PCR sputum testing, the accuracy of detecting COVID-19 from CXR images is lacking in the existing deep learning models. OBJECTIVE: This study aims to classify COVID-19 and normal patients from CXR images using semantic segmentation networks for detecting and labeling COVID-19 infected lung lobes in CXR images. METHODS: For semantically segmenting infected lung lobes in CXR images for COVID-19 early detection, three structurally different deep learning (DL) networks such as SegNet, U-Net and hybrid CNN with SegNet plus U-Net, are proposed and investigated. Further, the optimized CXR image semantic segmentation networks such as GWO SegNet, GWO U-Net, and GWO hybrid CNN are developed with the grey wolf optimization (GWO) algorithm. The proposed DL networks are trained, tested, and validated without and with optimization on the openly available dataset that contains 2,572 COVID-19 CXR images including 2,174 training images and 398 testing images. The DL networks and their GWO optimized networks are also compared with other state-of-the-art models used to detect COVID-19 CXR images. RESULTS: All optimized CXR image semantic segmentation networks for COVID-19 image detection developed in this study achieved detection accuracy higher than 92%. The result shows the superiority of optimized SegNet in segmenting COVID-19 infected lung lobes and classifying with an accuracy of 98.08% compared to optimized U-Net and hybrid CNN. CONCLUSION: The optimized DL networks has potential to be utilised to more objectively and accurately identify COVID-19 disease using semantic segmentation of COVID-19 CXR images of the lungs.

UBNet: Deep learning-based approach for automatic X-ray image detection of pneumonia and COVID-19 patients.

BACKGROUND: Analysis of chest X-ray images is one of the primary standards in diagnosing patients with COVID-19 and pneumonia, which is faster than using PCR Swab method. However, accuracy of using X-ray images needs to be improved. OBJECTIVE: To develop a new deep learning system of chest X-ray images and evaluate whether it can quickly and accurately detect pneumonia and COVID-19 patients. METHODS: The developed deep learning system (UBNet v3) uses three architectural hierarchies, namely first, to build an architecture containing 7 convolution layers and 3 ANN layers (UBNet v1) to classify between normal images and pneumonia images. Second, using 4 layers of convolution and 3 layers of ANN (UBNet v2) to classify between bacterial and viral pneumonia images. Third, using UBNet v1 to classify between pneumonia virus images and COVID-19 virus infected images. An open-source database with 9,250 chest X-ray images including 3,592 COVID-19 images were used in this study to train and test the developed deep learning models. RESULTS: CNN architecture with a hierarchical scheme developed in UBNet v3 using a simple architecture yielded following performance indices to detect chest X-ray images of COVID-19 patients namely, 99.6%accuracy, 99.7%precision, 99.7%sensitivity, 99.1%specificity, and F1 score of 99.74%. A desktop GUI-based monitoring and classification system supported by a simple CNN architecture can process each chest X-ray image to detect and classify COVID-19 image with an average time of 1.21 seconds. CONCLUSION: Using three hierarchical architectures in UBNet v3 improves system performance in classifying chest X-ray images of pneumonia and COVID-19 patients. A simple architecture also speeds up image processing time.

Predictive Modeling of Covid-19 Data in the US: Adaptive Phase-Space Approach.

There are currently intensified efforts by the scientific community world-wide to analyze the dynamics of the Covid-19 pandemic in order to predict key epidemiological effects and assist the proper planning for its clinical management, as well as guide sociopolitical decision-making regarding proper mitigation measures. Most efforts follow variants of the established SIR methodological framework that divides a population into "Susceptible", "Infectious" and "Recovered/Removed" fractions and defines their dynamic inter-relationships with first-order differential equations. GOAL: This paper proposes a novel approach based on data-guided detection and concatenation of infection waves - each of them described by a Riccati equation with adaptively estimated parameters. METHODS: This approach was applied to Covid-19 daily time-series data of US confirmed cases, resulting in the decomposition of the epidemic time-course into five "Riccati modules" representing major infection waves to date (June 18th). RESULTS: Four waves have passed the time-point of peak infection rate, with the fifth expected to peak on July 20th. The obtained parameter estimates indicate gradual reduction of infectivity rate, although the latest wave is expected to be the largest. CONCLUSIONS: This analysis suggests that, if no new waves of infection emerge, the Covid-19 epidemic will be controlled in the US (<5000 new daily cases) by September 26th, and the maximum of confirmed cases will reach 4,160,000. Importantly, this approach can be used to detect (via rigorous statistical methods) the emergence of possible new waves of infections in the future. Analysis of data from individual states or countries may quantify the distinct effects of different mitigation measures.

Peptide substrate screening for the diagnosis of SARS-CoV-2 using fluorescence resonance energy transfer (FRET) assay.

The novel corona (SARS-CoV-2) virus causes a global pandemic, which motivates researchers to develop reliable and effective methods for screening and detection of SARS-CoV-2. Though there are several methods available for the diagnosis of SARS-CoV-2 such as RT-PCR and ELSIA, nevertheless, these methods are time-consuming and may not apply at the point of care. In this study, we have developed a specific, sensitive, quantitative and fast detection method for SARS-CoV-2 by fluorescence resonance energy transfer (FRET) assay. The total extracellular protease proteolytic activity from the virus has been used as the biomarker. The specific peptide sequences from the library of 115 dipeptides were identified via changes in the fluorescence signal. The fluorogenic dipeptide substrates have the fluorophore and a quencher at the N- and the C- terminals, respectively. When the protease hydrolyzes the peptide bond between the two specific amino acids, it leads to a significant increase in the fluorescence signals. The specific fluorogenic peptide (H-d) produces a high fluorescence signal. A calibration plot was obtained from the changes in the fluorescence intensity against the different concentrations of the viral protease. The lowest limit of detection of this method was 9.7 ± 3 pfu/mL. The cross-reactivity of the SARS-CoV-2-specific peptide was tested against the MERS-CoV which does not affect the fluorescence signal. A significant change in the fluorescence signal with patient samples indicates that this FRET-based assay might be applied for the diagnosis of SARS-CoV-2 patients. Graphical abstract.

Detection of coronavirus disease from X-ray images using deep learning and transfer learning algorithms.

OBJECTIVE: This study aims to employ the advantages of computer vision and medical image analysis to develop an automated model that has the clinical potential for early detection of novel coronavirus (COVID-19) infected disease. METHOD: This study applied transfer learning method to develop deep learning models for detecting COVID-19 disease. Three existing state-of-the-art deep learning models namely, Inception ResNetV2, InceptionNetV3 and NASNetLarge, were selected and fine-tuned to automatically detect and diagnose COVID-19 disease using chest X-ray images. A dataset involving 850 images with the confirmed COVID-19 disease, 500 images of community-acquired (non-COVID-19) pneumonia cases and 915 normal chest X-ray images was used in this study. RESULTS: Among the three models, InceptionNetV3 yielded the best performance with accuracy levels of 98.63% and 99.02% with and without using data augmentation in model training, respectively. All the performed networks tend to overfitting (with high training accuracy) when data augmentation is not used, this is due to the limited amount of image data used for training and validation. CONCLUSION: This study demonstrated that a deep transfer learning is feasible to detect COVID-19 disease automatically from chest X-ray by training the learning model with chest X-ray images mixed with COVID-19 patients, other pneumonia affected patients and people with healthy lungs, which may help doctors more effectively make their clinical decisions. The study also gives an insight to how transfer learning was used to automatically detect the COVID-19 disease. In future studies, as the amount of available dataset increases, different convolution neutral network models could be designed to achieve the goal more efficiently.