Your browser doesn't support javascript.
Show: 20 | 50 | 100
Results 1 - 20 de 145
Filter
1.
21st International Conference on New Trends in Intelligent Software Methodologies, Tools and Techniques, SoMeT 2022 ; JOUR:596-606, 355.
Article in English | Scopus | ID: covidwho-2089732

ABSTRACT

This paper proposes a new deep learning model to detect COVID-19 lesions in chest CT images. This method is based on the Attention U-net which uses the layer of Atrous Spatial Pyramid Pooling (ASPP) to capture the feature on various scales. It also contains an attention gate. The attention gate provides the ability to suppress irrelevant regions and focus on the useful feature in an input image. The experimental results show that this method can achieve 99.61% accuracy and 80.43% precision. They are more effectively than the baseline method on Chest CT images. © 2022 The authors and IOS Press. All rights reserved.

2.
Microbiol Spectr ; : e0214322, 2022 Oct 26.
Article in English | MEDLINE | ID: covidwho-2088440

ABSTRACT

The pandemic of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has posed an enormous burden on the global public health system and has had disastrous socioeconomic consequences. Currently, single sampling tests, 20-in-1 pooling tests, nucleic acid point-of-care tests (POCTs), and rapid antigen tests are implemented in different scenarios to detect SARS-CoV-2, but a comprehensive evaluation of them is scarce and remains to be explored. In this study, 3 SARS-CoV-2 inactivated cell culture supernatants were used to evaluate the analytical performance of these strategies. Additionally, 5 recombinant SARS-CoV-2 nucleocapsid (N) proteins were also used for rapid antigen tests. For the wild-type (WT), Delta, and Omicron strains, the lowest inactivated virus concentrations to achieve 100% detection rates of single sampling tests ranged between 1.28 × 102 to 1.02 × 103, 1.28 × 102 to 4.10 × 103, and 1.28 × 102 to 2.05 × 103 copies/mL. The 20-in-1 pooling tests ranged between 1.30 × 102 to 1.04 × 103, 5.19 × 102 to 2.07 × 103, and 2.59 × 102 to 1.04 × 103 copies/mL. The nucleic acid POCTs were all 1.42 × 103 copies/mL. The rapid antigen tests ranged between 2.84 × 105 to 7.14 × 106, 8.68 × 104 to 7.14 × 106, and 1.12 × 105 to 3.57 × 106 copies/mL. For the WT, Delta AY.2, Delta AY.1/AY.3, Omicron BA.1, and Omicron BA.2 recombinant N proteins, the lowest concentrations to achieve 100% detection rates of rapid antigen tests ranged between 3.47 to 142.86, 1.74 to 142.86, 3.47 to 142.86, 3.47 to 142.86, and 5.68-142.86 ng/mL, respectively. This study provided helpful insights into the scientific deployment of tests and recommended the full-scale consideration of the testing purpose, resource availability, cost performance, result rapidity, and accuracy to facilitate a profound pathway toward the long-term surveillance of coronavirus disease 2019 (COVID-19). IMPORTANCE In the study, we reported an evaluation of 4 detection strategies implemented in different scenarios for SARS-CoV-2 detection: single sampling tests, 20-in-1 pooling tests, nucleic acid point-of-care tests, and rapid antigen tests. 3 SARS-CoV-2-inactivated SARS-CoV-2 cell culture supernatants and 5 recombinant SARS-CoV-2 nucleocapsid proteins were used for evaluation. In this analysis, we found that for the WT, Delta, and Omicron supernatants, the lowest concentrations to achieve 100% detection rates of single sampling tests ranged between 1.28 × 102 to 1.02 × 103, 1.28 × 102 to 4.10 × 103, and 1.28 × 102 to 2.05 × 103 copies/mL. The 20-in-1 pooling tests ranged between 1.30 × 102 to 1.04 × 103, 5.19 × 102 to 2.07 × 103, and 2.59 × 102 to 1.04 × 103 copies/mL. The nucleic acid POCTs were all 1.42 × 103 copies/mL. The rapid antigen tests ranged between 2.84 × 105 to 7.14 × 106, 8.68 × 104 to 7.14 × 106, and 1.12 × 105 to 3.57 × 106 copies/mL. For the WT, Delta AY.2, Delta AY.1/AY.3, Omicron BA.1, and Omicron BA.2 recombinant N proteins, the lowest concentrations to achieve 100% detection rates of rapid antigen tests ranged between 3.47 to 142.86, 1.74 to 142.86, 3.47 to 142.86, 3.47 to 142.86, and 5.68 to 142.86 ng/mL, respectively.

3.
Computers in Biology and Medicine ; : 106195, 2022.
Article in English | ScienceDirect | ID: covidwho-2068841

ABSTRACT

According to the World Health Organization, an estimate of more than five million infections and 355,000 deaths have been recorded worldwide since the emergence of the coronavirus disease (COVID-19). Various researchers have developed interesting and effective deep learning frameworks to tackle this disease. However, poor feature extraction from the Chest X-ray images and the high computational cost of the available models impose difficulties to an accurate and fast Covid-19 detection framework. Thus, the major purpose of this study is to offer an accurate and efficient approach for extracting COVID-19 features from chest X-rays that is also less computationally expensive than earlier research. To achieve the specified goal, we explored the Inception V3 deep artificial neural network. This study proposed LCSB-Inception;a two-path (L and AB channel) Inception V3 network along the first three convolutional layers. The RGB input image is first transformed to CIE LAB coordinates (L channel which is aimed at learning the textural and edge features of the Chest X-Ray and AB channel which is aimed at learning the color variations of the Chest X-ray images). The L achromatic channel and the AB channels filters are set to 50%L-50%AB. This method saves between one-third and one-half of the parameters in the divided branches. We further introduced a global second-order pooling at the last two convolutional blocks for more robust image feature extraction against the conventional max-pooling. The detection accuracy of the LCSB-Inception is further improved by employing the Contrast Limited Adaptive Histogram Equalization (CLAHE) image enhancement technique on the input image before feeding them to the network. The proposed LCSB-Inception network is experimented on using two loss functions (Categorically smooth loss and categorically Cross-entropy) and two learning rates whereas Accuracy, Precision, Sensitivity, Specificity F1-Score, and AUC Score were used for evaluation via the chestX-ray-15k (Data_1) and COVID-19 Radiography dataset (Data_2). The proposed models produced an acceptable outcome with an accuracy of 0.97867 (Data_1) and 0.98199 (Data_2) according to the experimental findings. In terms of COVID-19 identification, the suggested models outperform conventional deep learning models and other state-of-the-art techniques presented in the literature based on the results.

4.
4th International Conference on Intelligent Control, Measurement and Signal Processing, ICMSP 2022 ; : 748-752, 2022.
Article in English | Scopus | ID: covidwho-2052014

ABSTRACT

In order to prevent COVID-19 effectively, non-contact body temperature measurement and human identification are required in public places, but face recognition based on visible light cannot meet the requirements. Therefore, this paper proposes a thermal imaging face recognition method based on temperature block feature extraction. Histogram equalization and median filter are used to preprocess the face image, and Sobel operator is used for face detection;Six dimensional features including temperature mean, standard deviation and adjacent difference are extracted from each temperature block in the average poolinged temperature matrix, and classified by max-correlation-coefficient method. The experimental results show that the recognition rate of this method is 6.1% higher than that of PCA method with the temperature block size of boldsymbol{2times 2}. When using the same hardware to execute the program, if the two recognition rates are very close, the average test time of the proposed method is 22.2% less than the one of deep learning models such as Alexnet. Furthermore, the proposed method has strong robustness for small training sample set. For example, the recognition rate of single training sample model can reach 0.7, while in the deep learning model, except Mobilenet can reach 0.6, all of the others are less than 0.4. © 2022 IEEE.

5.
IEEE Sens J ; 22(18): 17573-17582, 2022 Sep.
Article in English | MEDLINE | ID: covidwho-2037820

ABSTRACT

(Aim) COVID-19 pandemic causes numerous death tolls till now. Chest CT is an effective imaging sensor system to make accurate diagnosis. (Method) This article proposed a novel seven layer convolutional neural network based smart diagnosis model for COVID-19 diagnosis (7L-CNN-CD). We proposed a 14-way data augmentation to enhance the training set, and introduced stochastic pooling to replace traditional pooling methods. (Results) The 10 runs of 10-fold cross validation experiment show that our 7L-CNN-CD approach achieves a sensitivity of 94.44±0.73, a specificity of 93.63±1.60, and an accuracy of 94.03±0.80. (Conclusion) Our proposed 7L-CNN-CD is effective in diagnosing COVID-19 in chest CT images. It gives better performance than several state-of-the-art algorithms. The data augmentation and stochastic pooling methods are proven to be effective.

6.
J Infect Dev Ctries ; 16(8): 1278-1284, 2022 08 30.
Article in English | MEDLINE | ID: covidwho-2030101

ABSTRACT

INTRODUCTION: Mass testing is essential in the surveillance strategy for fighting the COVID-19 pandemic. It allows early detection of suspected cases and subsequently early isolation to mitigate spread. However, the high cost and limited consumables and reagents hinder the mass testing strategy in developing countries such as Indonesia. The specimen pooling strategy is an option to perform mass screening with limited resources. This study aims to determine the positivity rate cut-off and to evaluate the efficiency of pooling strategy for the laboratory diagnosis of COVID-19. METHODOLOGY: Between August 4th, 2020, and November 11th, 2020, a four-sample pooling strategy testing to detect SARS-CoV-2 was carried out at the Microbiology Diagnostic Laboratory of Diponegoro National Hospital, Semarang, Indonesia. Pools with positive results were subjected to individual specimen retesting. Spearman's correlation and linear regression analysis were used to determine the best positivity rate cut-off to apply pooling strategy. RESULTS: A total of 15,216 individual specimens were pooled into 3,804 four-sample pools. Among these pools, 1,007 (26.47%) were positive. Five hundred and ten (50.64%) were 1/4 positive. A maximum positivity rate of 22% is needed to save at least 50% extraction and qRT-PCR reactions in a four-sample pooling strategy. CT values between individual specimens and pools showed a good interval agreement. CONCLUSIONS: Pooling strategy could reduce personnel workload and reagent cost, and increase laboratory capacity by up to 50% when the positivity rate is less than 22%.


Subject(s)
COVID-19 , COVID-19/diagnosis , COVID-19 Testing , Humans , Pandemics , SARS-CoV-2 , Sensitivity and Specificity , Specimen Handling/methods
7.
Math Biosci Eng ; 19(11): 11018-11033, 2022 08 01.
Article in English | MEDLINE | ID: covidwho-2024422

ABSTRACT

Various measures have been implemented around the world to prevent the spread of SARS-CoV-2. A potential tool to reduce disease transmission is regular mass testing of a high percentage of the population, possibly with pooling (testing a compound of several samples with one single test). We develop a compartmental model to study the applicability of this method and compare different pooling strategies: regular and Dorfman pooling. The model includes isolated compartments as well, from where individuals rejoin the active population after some time delay. We develop a method to optimize Dorfman pooling depending on disease prevalence and establish an adaptive strategy to select variable pool sizes during the course of the epidemic. It is shown that optimizing the pool size can avert a significant number of infections. The adaptive strategy is much more efficient, and may prevent an epidemic outbreak even in situations when a fixed pool size strategy can not.


Subject(s)
COVID-19 , COVID-19/epidemiology , Disease Outbreaks , Epidemiological Models , Humans , Prevalence , SARS-CoV-2
8.
14th IEEE International Conference on Signal Processing and Communications, SPCOM 2022 ; 2022.
Article in English | Scopus | ID: covidwho-2018987

ABSTRACT

Computed Tomography (CT) based analysis will assist doctors in a prompt diagnosis of the Covid-19 infection. Automated segmentation of lesions in chest CT scans helps in determining the severity of the infection. The presented work addresses the task of automated segmentation of Covid-19 lesions. A U-Net framework incorporated with spatial-channel attention modules (contextual relationships), Atrous Spatial Pyramid Pooling module (a wider receptive field) and Deep Supervision (lesion focus, less error propagation) is proposed. Focal Tversky Loss is used to evaluate the outputs at coarser scales while Tversky loss evaluates the final segmentation output. This combination of losses is used to enhance segmentation of the small lesions. The framework is trained on CT scans of 20 subjects of COVID19 CT Lung and Infection Segmentation Dataset and tested on Mosmed dataset of 50 subjects, where infection has affected less than 25% of lung parenchyma. The experimental results show that the proposed method is effective in segmenting the hard ROIs in Mosmed data resulting in a mean Dice score of 0.57 (9% more than the state-of-the-art). © 2022 IEEE.

9.
Int J Environ Res Public Health ; 19(17)2022 Sep 02.
Article in English | MEDLINE | ID: covidwho-2010032

ABSTRACT

The conversion rate between asymptomatic infections and reported/unreported symptomatic infections is a very sensitive parameter for model variables that spread COVID-19. This is important information for follow-up use in screening, prediction, prognostics, contact tracing, and drug development for the COVID-19 pandemic. The model described here suggests that there may not be enough researchers to solve all of these problems thoroughly and effectively, and it requires careful selection of what we are doing and rapid sharing of results and models and optimizing modeling simulations with value to reduce the impact of COVID-19. Exploring simulation modeling will help decision makers make the most informed decisions. In order to fight against the "Delta" virus, the establishment of a line of defense through all-people testing (APT) is not only an effective method summarized from past experience but also one of the best means to effectively cut the chain of epidemic transmission. The effect of large-scale testing has been fully verified in the international community. We developed a practical dynamic infectious disease model-SETPG (A + I) RD + APT by considering the effects of the all-people test (APT). The model is useful for studying effects of screening measures and providing a more realistic modelling with all-people-test strategies, which require everybody in a population to be tested for infection. In prior work, a total of 370 epidemic cases were collected. We collected three kinds of known cases: the cumulative number of daily incidences, daily cumulative recovery, and daily cumulative deaths in Hong Kong and the United States between 22 January 2020 and 13 November 2020 were simulated. In two essential strategies of the integrated SETPG (A + I) RD + APT model, comparing the cumulative number of screenings in derivative experiments based on daily detection capability and tracking system application rate, we evaluated the performance of the timespan required for the basic regeneration number (R0) and real-time regeneration number (R0t) to reach 1; the optimal policy of each experiment is available, and the screening effect is evaluated by screening performance indicators. with the binary encoding screening method, the number of screenings for the target population is 8667 in HK and 1,803,400 in the U.S., including 6067 asymptomatic cases in HK and 1,262,380 in the U.S. as well as 2599 cases of mild symptoms in HK and 541,020 in the U.S.; there were also 8.25 days of screening timespan in HK and 9.25 days of screening timespan required in the U.S. and a daily detectability of 625,000 cases in HK and 6,050,000 cases in the U.S. Using precise tracking technology, number of screenings for the target population is 6060 cases in HK and 1,766,420 cases in the U.S., including 4242 asymptomatic cases in HK and 1,236,494 cases in the U.S. as well as 1818 cases of mild symptoms in HK and 529,926 cases in the U.S. Total screening timespan (TS) is 8.25~9.25 days. According to the proposed infectious dynamics model that adapts to the all-people test, all of the epidemic cases were reported for fitting, and the result seemed more reasonable, and epidemic prediction became more accurate. It adapted to densely populated metropolises for APT on prevention.


Subject(s)
COVID-19 , Communicable Diseases , COVID-19/diagnosis , COVID-19/epidemiology , COVID-19 Testing , Communicable Diseases/epidemiology , Humans , Pandemics/prevention & control , SARS-CoV-2 , United States
10.
Front Microbiol ; 13: 858555, 2022.
Article in English | MEDLINE | ID: covidwho-1997458

ABSTRACT

An effective and rapid diagnosis has great importance in tackling the ongoing COVID-19 pandemic through isolation of the infected individuals to curb the transmission and initiation of specialized treatment for the disease. It has been proven that enhanced testing capacities contribute to efficiently curbing SARS-CoV-2 transmission during the initial phases of the outbreaks. RT-qPCR is considered a gold standard for the diagnosis of COVID-19. However, in resource-limited countries expenses for molecular diagnosis limits the diagnostic capacities. Here, we present interventions of two pooling strategies as 5 sample pooling (P-5) and 10 sample pooling (P-10) in a high-throughput COVID-19 diagnostic laboratory to enhance throughput and save resources and time over a period of 6 months. The diagnostic capacity was scaled-up 2.15-folds in P-5 and 1.8-fold in P-10, reagents (toward RNA extraction and RT-qPCR) were preserved at 75.24% in P-5 and 86.21% in P-10, and time saved was 6,290.93 h in P-5 and 3147.3 h in P-10.

11.
7th IEEE International conference for Convergence in Technology, I2CT 2022 ; 2022.
Article in English | Scopus | ID: covidwho-1992600

ABSTRACT

Since December 2019, the world is fighting against the newly found virus named COVID-19 whose symptoms are closer to pneumonia. Being highly contagious, it has spread all over the world, and hence the World Health Organization has declared this as a global pandemic. Some patients infected with this virus have severe symptoms which are fatal. Hence the early discovery of COVID-19 infected patients is necessary to avoid further community spread. The available tests such as RTPCR and Rapid Antigen Tests are not 100% accurate and do not give quick results either. Therefore, it is the need of the hour to explore identification methodologies that are quick, accurate, and easily scalable. This work intends to do so using different machine learning and deep learning models. First, the step involves feature extraction using Gray Level Co-occurrence Matrix (GLCM) and classification with LightGBM classifier which gives an accuracy of 92.78%. This is then further improved to 95.79% using wavelets. Further, the CNN architectures with max-pooling and DWT layers are compared and it's found that CNN architecture with max-pooling layer gives better accuracy of 95.72%. Thus, this work presents a comparative analysis of Machine Learning Algorithms and CNN architectures for better accuracy and time. © 2022 IEEE.

12.
Front Microbiol ; 13: 957957, 2022.
Article in English | MEDLINE | ID: covidwho-1987529

ABSTRACT

COVID-19 is a life-threatening multisistemic infection caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Infection control relies on timely identification and isolation of infected people who can alberg the virus for up to 14 days, providing important opportunities for undetected transmission. This note describes the application of rRT-PCR test for simpler, faster and less invasive monitoring of SARS-CoV-2 infection using pooling strategy of samples. Seventeen positive patients were provided with sterile dry swabs and asked to self-collected 2 nasal specimens (#NS1 and #NS2). The #NS1 was individually placed in a single tube and the #NS2 was placed in another tube together with 19 NSs collected from 19 negative patients. Both tubes were then tested with conventional molecular rRT-PCR and the strength of pooling nasal testing was compared with the molecular test performed on the single NS of each positive patient. The pooling strategy detected SARS-CoV-2 RNA to a similar extent to the single test, even when Ct value is on average high (Ct 37-38), confirming that test sensibility is not substantially affected even if the pool contains only one low viral load positive sample. Furthermore, the pooling strategy have benefits for SARS-CoV-2 routinary monitoring of groups in regions with a low SARS-CoV-2 prevalence.

13.
View ; 3(4), 2022.
Article in English | ProQuest Central | ID: covidwho-1958862

ABSTRACT

The outbreak of coronavirus disease 2019 (COVID‐19, caused by SARS‐Cov‐2) is a big challenge for global health systems and the economy. Rapid and accurate tests are crucial at early stages of this pandemic. Reverse transcription‐quantitative real‐time polymerase chain reaction is the current gold standard method for detection of SARS‐Cov‐2. It is impractical and costly to test individuals in large‐scale population screens, especially in low‐ and middle‐income countries due to their shortage of nucleic acid testing reagents and skilled staff. Accordingly, sample pooling, such as for blood screening for syphilis, is now widely applied to COVID‐19. In this paper, we survey and review several different pooled‐sample testing strategies, based on their group size, prevalence, testing number, and sensitivity, and we discuss their efficiency in terms of reducing cost and saving time while ensuring sensitivity.

14.
Int J Data Sci Anal ; : 1-21, 2022 May 06.
Article in English | MEDLINE | ID: covidwho-1943732

ABSTRACT

Conspiracy theories have seen a rise in popularity in recent years. Spreading quickly through social media, their disruptive effect can lead to a biased public view on policy decisions and events. We present a novel approach for LDA-pre-processing called Iterative Filtering to study such phenomena based on Twitter data. In combination with Hashtag Pooling as an additional pre-processing step, we are able to achieve a coherent framing of the discussion and topics of interest, despite of the inherent noisiness and sparseness of Twitter data. Our novel approach enables researchers to gain detailed insights into discourses of interest on Twitter, allowing them to identify tweets iteratively that are related to an investigated topic of interest. As an application, we study the dynamics of conspiracy-related topics on US Twitter during the last four months of 2020, which were dominated by the US-Presidential Elections and Covid-19. We monitor the public discourse in the USA with geo-spatial Twitter data to identify conspiracy-related contents by estimating Latent Dirichlet Allocation (LDA) Topic Models. We find that in this period, usual conspiracy-related topics played a marginal role in comparison with dominating topics, such as the US-Presidential Elections or the general discussions about Covid-19. The main conspiracy theories in this period were the ones linked to "Election Fraud" and the "Covid-19-hoax." Conspiracy-related keywords tended to appear together with Trump-related words and words related to his presidential campaign.

15.
PeerJ ; 10: e13277, 2022.
Article in English | MEDLINE | ID: covidwho-1934568

ABSTRACT

Importance: The rise of novel, more infectious SARS-CoV-2 variants has made clear the need to rapidly deploy large-scale testing for COVID-19 to protect public health. However, testing remains limited due to shortages of personal protective equipment (PPE), naso- and oropharyngeal swabs, and healthcare workers. Simple test methods are needed to enhance COVID-19 screening. Here, we describe a simple, and inexpensive spit-test for COVID-19 screening called Patient Self-Collection of Sample-CoV2 (PSCS-CoV2). Objective: To evaluate an affordable and convenient test for COVID-19. Methods: The collection method relies on deep throat sputum (DTS) self-collected by the subject without the use of swabs, and was hence termed the Self-Collection of Sample for SARS-CoV-2 (abbreviated PSCS-CoV2). We used a phenol-chloroform extraction method for the viral RNA. We then tested for SARS-CoV-2 using real-time reverse transcription polymerase chain reaction with primers against at least two coding regions of the viral nucleocapsid protein (N1 and N2 or E) of SARS-CoV-2. We evaluted the sensitivity and specificity of our protocol. In addition we assess the limit of detection, and efficacy of our Viral Inactivating Solution. We also evaluated our protocol, and pooling strategy from volunteers on a local college campus. Results: We show that the PSCS-CoV2 method accurately identified 42 confirmed COVID-19 positives, which were confirmed through the nasopharyngeal swabbing method of an FDA approved testing facility. For samples negative for COVID-19, we show that the cycle threshold for N1, N2, and RP are similar between the PSCS-CoV2 and nasopharynx swab collection method (n = 30). We found a sensitivity of 100% (95% Confidence Interval [CI], 92-100) and specifity of 100% (95% CI, 89-100) for our PSCS-CoV2 method. We determined our protocol has a limit of detection of 1/10,000 for DTS from a COVID-19 patient. In addition, we show field data of the PSCS-CoV2 method on a college campus. Ten of the twelve volunteers (N1 < 30) that we tested as positive were subsequently tested positive by an independent laboratory. Finally, we show proof of concept of a pooling strategy to test for COVID-19, and recommend pool sizes of four if the positivity rate is less than 15%. Conclusion and Relevance: We developed a DTS-based protocol for COVID-19 testing with high sensitivity and specificity. This protocol can be used by non-debilitated adults without the assistance of another adult, or by non-debilitated children with the assistance of a parent or guardian. We also discuss pooling strategies based on estimated positivity rates to help conserve resources, time, and increase throughput. The PSCS-CoV2 method can be a key component of community-wide efforts to slow the spread of COVID-19.

16.
IEEE Internet of Things Journal ; 9(13):11376-11384, 2022.
Article in English | Scopus | ID: covidwho-1932130

ABSTRACT

Up to now, the coronavirus disease 2019 (COVID-19) has been sweeping across all over the world, which has affected individual's lives in an overwhelming way. To fight efficiently against the COVID-19, radiography and radiology images are used by clinicians in hospitals. This article presents an integrated framework, named COVIDNet, for classifying COVID-19 patients and healthy controls. Specifically, ResNet (i.e., ResNet-18 and ResNet-50) is adopted as a backbone network to extract the discriminative features first. Second, the spatial pyramid pooling (SPP) layer is adopted to capture the middle-level features from the features of ResNet. To learn the high-level features, the NetVLAD layer is used to aggregate the features representation from middle-level features. The context gating (CG) mechanism is adopted to further learn the high-level features for predicting the COVID-19 patients or not. Finally, extensive experiments are conducted on the collected database, showing the excellent performance of the proposed integrated architecture, with the sensitivity up to 97% and specificity of 99.5% of the ResNet-18, and with the sensitivity up to 99% and specificity of 99.4% of the ResNet-50. © 2014 IEEE.

17.
Rev Esp Quimioter ; 35(4): 401-405, 2022 Aug.
Article in English | MEDLINE | ID: covidwho-1904218

ABSTRACT

OBJECTIVE: Since the first cases of SARS-CoV-2 appeared, there have been numerous techniques that have been developed for the diagnosis or monitoring of infection, both direct and serological techniques. Choosing a good diagnostic tool is essential for epidemiological control. The objective was to compare five commercialized RT-PCR techniques in real time, in sensitivity, specificity and agreement for the detection of SARS-CoV-2. METHODS: Five commercial RT-PCR kits for the detection of SARS-CoV-2 were compared. Eight known positive samples were taken and subjected to seven different dilutions or concentrations, and another 135 negative samples were used to determine sensitivity, specificity, and agreement values. RESULTS: The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for the Palex, Roche and GeneXpert techniques with respect to Seegene were identical, corresponding to 98.21%, 100%, 100% and 99.26% respectively. For Becton Dickinson the sensitivity was 89.28%, the specificity of 100%, the PPV of 100% and the NPV of 95.74%. The agreement using the Kappa index for Palex, Roche and GeneXpert was 0.9892, while the agreement for Becton Dickinson was with a Kappa index of 0.9215. CONCLUSIONS: All commercial RT-PCR kits had high sensitivities and specificities, as well as PPV, NPV, and concordance.


Subject(s)
COVID-19 , SARS-CoV-2 , COVID-19/diagnosis , COVID-19 Testing , Humans , Polymerase Chain Reaction , SARS-CoV-2/genetics , Sensitivity and Specificity
18.
BMC Med Res Methodol ; 22(1): 148, 2022 05 21.
Article in English | MEDLINE | ID: covidwho-1902354

ABSTRACT

BACKGROUND: Missing data prove troublesome in data analysis; at best they reduce a study's statistical power and at worst they induce bias in parameter estimates. Multiple imputation via chained equations is a popular technique for dealing with missing data. However, techniques for combining and pooling results from fitted generalized additive models (GAMs) after multiple imputation have not been well explored. METHODS: We simulated missing data under MCAR, MAR, and MNAR frameworks and utilized random forest and predictive mean matching imputation to investigate a variety of rules for combining GAMs after multiple imputation with binary and normally distributed outcomes. We compared multiple pooling procedures including the "D2" method, the Cauchy combination test, and the median p-value (MPV) rule. The MPV rule involves simply computing and reporting the median p-value across all imputations. Other ad hoc methods such as a mean p-value rule and a single imputation method are investigated. The viability of these methods in pooling results from B-splines is also examined for normal outcomes. An application of these various pooling techniques is then performed on two case studies, one which examines the effect of elevation on a six-minute walk distance (a normal outcome) for patients with pulmonary arterial hypertension, and the other which examines risk factors for intubation in hospitalized COVID-19 patients (a dichotomous outcome). RESULTS: In comparison to the results from generalized additive models fit on full datasets, the median p-value rule performs as well as if not better than the other methods examined. In situations where the alternative hypothesis is true, the Cauchy combination test appears overpowered and alternative methods appear underpowered, while the median p-value rule yields results similar to those from analyses of complete data. CONCLUSIONS: For pooling results after fitting GAMs to multiply imputed datasets, the median p-value is a simple yet useful approach which balances both power to detect important associations and control of Type I errors.


Subject(s)
COVID-19 , Hypertension, Pulmonary , COVID-19/epidemiology , Colorado , Hospitalization , Humans , Hypertension, Pulmonary/diagnosis , Models, Statistical , Registries
19.
47th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2022 ; 2022-May:4518-4522, 2022.
Article in English | Scopus | ID: covidwho-1891397

ABSTRACT

The ongoing pandemic and the necessity of frequent testing have spurred a growing interest in pooled testing. Conventional recovery methods from pooled tests are based on group testing or compressed sensing tools which rely on simplistic modeling of the pooling process, and may not be reliable in the presence of complex and noisy measurement procedures and highly infected populations. In this work, we propose a strategy for pooled testing designed for noisy settings, which bypasses the need for a tractable acquisition model. This is achieved by combining deep learning, for implicitly learning the measurement relationship from data, with factor graph inference, which exploits the structured known pooling pattern. Learned factor graphs provide a quantitative readout corresponding to the infection severity, as opposed to group testing which only detects the presence of infection. The proposed scheme is shown to achieve improved robustness to noise compared with previous approaches and to reliably estimate in highly infected populations. © 2022 IEEE

20.
Ieee Transactions on Emerging Topics in Computational Intelligence ; : 10, 2022.
Article in English | Web of Science | ID: covidwho-1886622

ABSTRACT

Coronavirus disease 2019 (COVID-19) generated a global public health emergency since December 2019, causing huge economic losses. To help radiologists strengthen their recognition of COVID-19 cases, we developed a computer-aided diagnosis system based on deep learning to automatically classify chest computed tomography-based COVID-19, Tuberculosis, and healthy control subjects. Our novel classification model AdaD-FNN sequentially transfers the trained knowledge of an FNN estimator to the next FNN estimator while updating the weights of the samples in the training set with a decaying learning rate. This model inhibits the network from remembering the noisy information and improves the learning of complex patterns in the hard-to-identify samples. Moreover, we designed a novel image preprocessing model F-U2MNet-C by enhancing the image features using fuzzy stacking and eliminating the interference factors using U2MNet segmentation. Extensive experiments are conducted on four publicly available datasets namely, TLDCA, UCSD-Al4H, SARS-CoV-2, TCIA, and the obtained classification accuracies are 99.52%, 92.96%, 97.86%, 91.97%. Our novel system gives out compelling performance for assisting COVID-19 detection when compared with 22 state-of-the-art methods. We hope to help link together biomedical research and artificial intelligence and to assist the diagnosis of doctors, radiologists, and inspectors at each epidemic prevention site in the real world.

SELECTION OF CITATIONS
SEARCH DETAIL