A Hybrid Model of Deep Learning Features and Clinical Features for Severe Cases Predication of COVID-19

COVID-19 has spread around the world since 2019. Approximately 6.5% of COVID-19 a risk of developing severe disease with high mortality rate. To reduce the mortality rate and provide appropriate treatment, this research established an integrated models with to predict the clinical outcome of COVID-19 patients with clinical, deep learning and radiomics features. To obtain the optimal feature combination for prediction, 9 clinical features combination was selected from all available clinical factors after using LASSO, 18 deep learning features from U-Net architecture, and 9 radiomics features from segmentation result. A total of 213 COVID-19 patients and 335 non-COVID-19 patients from 5 hospitals were enrolled and used as training and test sample in this research. The proposed model obtained an accuracy, precision, recall, specificity, F1-score and ROC curve of 0.971, 0.943, 0.937, 0.974, 0.941 and 0.979, respectively, which exceeds the related work using only clinical, deep learning or radiomics factors. © 2023 SPIE.

[Estimation of COVID-19 incidence in Shanghai under optimized epidemic prevention and control strategies].

Objective: To quantitatively estimate the incidence of COVID-19 in different backgrounds, including vaccination coverage, non-pharmacological interventions (NPIs) measures, home quarantine willingness and international arrivals, and the demands of healthcare resource in Shanghai in the context of optimized epidemic prevention and control strategies. Methods: Based on the natural history of 2019-nCoV, local vaccination coverage and NPI performance, an age-structured Susceptible-Exposed-Infections-Removed (SEIR) epidemic dynamic model was established for the estimation of the incidence of COVID-19 and demand of hospital beds in Shanghai by using the data on December 1, 2022 as the basis. Results: Based on current vaccination coverage, it is estimated that 180 184 COVID-19 cases would need treatment in hospitals in Shanghai within 100 days. When the booster vaccination coverage reaches an ideal level, the number of the cases needing hospitalization would decrease by 73.20%. School closure or school closure plus workplace closure could reduce the peak demand of regular beds by 24.04% or 37.73%, respectively, compared with the situation without NPI. Increased willingness of home quarantine could reduce the number of daily new cases and delay incidence peak of COVID-19. The number of international arrivals has little impact on the development of the epidemic. Conclusions: According to the epidemiological characteristics of COVID-19 and the actual situation of vaccination in Shanghai, the incidence of COVID-19 and health resource demand might be reduced by increasing vaccination coverage and early implementation of NPI.

Cost analysis of launching telemedicine services by Peking union Medicalcollege hospital during the pandemic of Coronavirus disease 2019.

Objective To assess the cost of launching telemedicine services by Peking Union Medical College Hospital (PUMCH) during coronavirus disease 2019 pandemic. Methods The patients using telemedicine services were enrolled during the period of pilot run from February 10th to April 15th, 2020. The study was done from the social perspective. A decision-tree model was constructed to compare the costs between telemedicine services and conventional clinical services for outpatients. The main outcome was measured as incre- mental cost-effective ness ratios (ICER). Sensitivity analysis was conducted by using one-way sensitivity analysis. Results During a period of forty-seven days, the online fever clinic was applied 3055 person-times(2070 patients) and the online outpatient clinic were applied 36 549 person-times(20 467 patients). On average, 44 febrile cases/d and 435 nonfebrile cases/d were reduced in the outpatient clinic. It helped to reduce roughly 1/4 (febrile) and 1/5(nonfebrile) of total numbers of the patients in the outpatient clinic during the peak period of the epidemic. If calculated according to the actual free-of-charge condition, the ICER was -64.7 yuans/person-time. If the actual cost of each consultant of telemedicine service was estimated according to the level of outpatient-service fee, the ICER was -5.5 yuans/person-time. The results of sensitivity analysis showed that the main factors affecting the ICERs were transportation cost, lost wages, and the efficiency of telemedicine services. Conclusions Launching telemedicine services helped to relieve the pressure at the outpatient clinics, and has the potential to provide significant cost saving compared to conventional clinic services for outpatients. It is worth considering applying this practice widely in the medical and health services.Copyright © 2021, Peking Union Medical College Hospital. All rights reserved.

Government Protection of Both Parties in the Operation of the Post-epidemic Labor Market in China

This chapter, "Government Protection of Both Parties in the Operation of the Post-Epidemic Labor Market in China”, considers the situation with regulation of the labor market in China at the COVID and post-COVID period. In the post-epidemic period, when the epidemic prevention and control initially achieved results, China, as the first country to discover and report COVID-19 and the most successful country in epidemic prevention and control, implemented government protection policies for both sides in the labor market in terms of stabilizing employment, which achieved remarkable results and served as a model for other countries. Generally, the government protection policies of the Chinese government for both enterprises and labors include the following aspects. First, monetary, fiscal, and employment policies work together to stabilize employment, strengthen protection for both sides, and ensure the sound operation of the labor market. Second, the forms of employment are standardized and diversified in accordance with the law to effectively protect the legitimate rights and interests of labors and give enterprises greater flexibility and convenience in employment. Third, special assistance is offered to enterprises in difficulty, key employment groups, and the unemployed. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2022.

Social determinants of health, prenatal maternal stress, and earlier birth during the COVID-19 pandemic

During the COVID-19 pandemic, pregnant women, especially those from socioeconomically disadvantaged and marginalized groups, experienced unprecedented stress. Prenatal stress and social determinants of health (SDoH) such as lower education and lack of a relationship partner are known to contribute to earlier birth. However, whether SDoH and stress independently contribute or whether the harmful impact of SDoH is mediated by stress is unknown. Moreover, the contributions of these factors has not been investigated in the context of a communal health crisis such as the COVID-19 pandemic. To examine these processes, we used a longitudinal cohort of 2473 women pregnant during the COVID-19 pandemic who reported a live birth. We compared structural equation models predicting gestational age at birth from SDoH (race/ethnicity, education, financial security, health insurance, relationship status, and lifetime abuse) and from prenatal maternal stress related and unrelated to the COVID-19 pandemic. Results indicate that the association of SDoH with earlier birth was partially mediated by prenatal stress. These findings help uncover mechanisms explaining health disparities in the U.S. and highlight the need to address both SDoH and the stress that these factors produce in under-resourced and marginalized communities. © 2023 John Wiley & Sons Ltd.

The IMU Augment SLAM on Unmanned Vehicle for Detection of Protective Measures in COVID-19

Detecting protective measures (e.g., masks, goggles and protective clothing) is a momentous step in the fight against COVID-19. The detection mode of unmanned devices based on Simultaneous localization and mapping (SLAM) and fusion technology is more efficient, economical and safe than the traditional manual detection. In this paper, a tightly-coupled nonlinear optimization approach is used to augment the visual feature extraction of SLAM by the gyroscope of the IMU to obtain a high-precision visual inertial system for joint position and pose estimation. Based on the VINS-Mono frame, first, an LSD algorithm based on a conditional selection strategy is proposed to extract line features efficiently. Then, we propose recovering missing point features from line features. Moreover, we propose a strategy to recover vanishing point features from line features, and add residuals to the SLAM cost function based on optimization, which optimizes point-line features in real time to promote the tracking and matching accuracy. Second, the wavelet threshold denoising method based on the 3σ criterion is used to carry out real-time online denoising for gyroscope to improve the output precision. Our WD-PL-VINS was measured on publicly available EuRoC datasets, TUM VI datasets and evaluated and validated in lab testing with a unmanned vehicle (UV) based on the NVIDIA Jetson-TX2 development board. The results show that our method's APE and RPE on MH-03-easy sequences are improved by 69.28% and 97.66%, respectively, compared with VINS-Mono. © 2001-2012 IEEE.

Fusion of Bilateral 2DPCA Information for Image Reconstruction and Recognition

Being an efficient image reconstruction and recognition algorithm, two-dimensional PCA (2DPCA) has an obvious disadvantage in that it treats the rows and columns of images unequally. To exploit the other lateral information of images, alternative 2DPCA (A2DPCA) and a series of bilateral 2DPCA algorithms have been proposed. This paper proposes a new algorithm named direct bilateral 2DPCA (DB2DPCA) by fusing bilateral information from images directly-that is, we concatenate the projection results of 2DPCA and A2DPCA together as the projection result of DB2DPCA and we average between the reconstruction results of 2DPCA and A2DPCA as the reconstruction result of DB2DPCA. The relationships between DB2DPCA and related algorithms are discussed under some extreme conditions when images are reshaped. To test the proposed algorithm, we conduct experiments of image reconstruction and recognition on two face databases, a handwritten character database and a palmprint database. The performances of different algorithms are evaluated by reconstruction errors and classification accuracies. Experimental results show that DB2DPCA generally outperforms competing algorithms both in image reconstruction and recognition. Additional experiments on reordered and reshaped databases further demonstrate the superiority of the proposed algorithm. In conclusion, DB2DPCA is a rather simple but highly effective algorithm for image reconstruction and recognition.

FGraDA: A Dataset and Benchmark for Fine-Grained Domain Adaptation in Machine Translation

Previous research for adapting a general neural machine translation (NMT) model into a specific domain usually neglects the diversity in translation within the same domain, which is a core problem for domain adaptation in real-world scenarios. One representative of such challenging scenarios is to deploy a translation system for a conference with a specific topic, e.g., global warming or coronavirus, where there are usually extremely less resources due to the limited schedule. To motivate wider investigation in such a scenario, we present a real-world fine-grained domain adaptation task in machine translation (FGraDA). The FGraDA dataset consists of Chinese-English translation task for four sub-domains of information technology: autonomous vehicles, AI education, real-time networks, and smart phone. Each sub-domain is equipped with a development set and test set for evaluation purposes. To be closer to reality, FGraDA does not employ any in-domain bilingual training data but provides bilingual dictionaries and wiki knowledge base, which can be easier obtained within a short time. We benchmark the fine-grained domain adaptation task and present in-depth analyses showing that there are still challenging problems to further improve the performance with heterogeneous resources. © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0.

[Impact of COVID-19 epidemic on syphilis case reporting in China].

Objective: To analyze the impact of COVID-19 epidemic on syphilis case reporting in China, and provide evidence to evaluate the epidemic situation of syphilis and strengthen the prevention and control of syphilis during COVID-19 epidemic. Methods: The data were collected from the National Notifiable Infectious Disease Reporting System of China Information System for Disease Control and Prevention, National STD Management Information System, and the "nCov2019" R package of github website. The changes of reported cases of syphilis before and during COVID-19 epidemic in China were analyzed. Joinpoint regression model was established by using the reported case number of syphilis from 2010 to 2018, the data in 2019 was used for validation, and the number of syphilis cases in 2020 and 2021 was predicted. The impact of COVID-19 epidemic on the number of syphilis cases was evaluated with calculating the percentage error (PE) between actual number and predicted number of syphilis cases reported.The correlation between reported cases of syphilis and COVID-19 was analyzed by Spearman's correlation analysis. The softwares of Joinpoint 4.9.1.0 and SPSS 18.0 were used for statistical analysis. Results: In 2020 and 2021, the reported cases of syphilis in China decreased significantly by 13.32% and 10.41%, respectively, compared with 2019 (before COVID-19 epidemic), and the reported cases of syphilis in 2021 increased by 3.36% compared with 2020. The reported cases of syphilis in 2020 and 2021 decreased by 17.95% and 20.41%, respectively, compared with predicted numbers. From January to March 2020, the reported monthly case number of syphilis was completely negatively correlated with the confirmed case number of COVID-19 (rs=-1.00, P<0.001). In the provinces with different scales of COVID-19 epidemic, there was also a negative correlation between the monthly reported case number of syphilis and confirmed case number of COVID-19 (all P<0.05). Conclusions: In China, the change of reported cases of syphilis was closely associated with COVID-19 epidemic in 2020 and 2021. Due to the influence of COVID-19 epidemic, the number of reported cases of syphilis decreased significantly, but it should not be thought that syphilis incidence will become a decline trend in the future. It is necessary to carefully and scientifically assess the changes in syphilis epidemic.

Security strategy of large general hospitals in the context of COVID-19 epidemic prevention and control

Coronavirus disease 2019 (COVID-19), characterized by high infectivity and invisibility, has spread quickly throughout the world and brought great challenges to hospital security work. Regarding the medical service reality of large general hospitals, the Security Department actively responded to the epidemic prevention and control work. They classified the risk areas in the hospital and formulated corresponding security strategies. Current paper summarize the improvement of the existing security management and control system, the management level of major public health emergencies and the emergency management. Then the control measures for scientific deployment of personnel and proper planning of the diagnosis and treatment process of risk patients during the epidemic are discussed. This study aimed to provide helpful reference for ensuring the stability of medical reception order in large general hospitals under COVID-19 epidemic. Copyright © 2022, Editorial Board of Journal of Xi'an Jiaotong University (Medical Sciences). All right reserved.

Preparation of monoclonal antibody against severe acute respiratory syndrome coronavirus 2 spike protein and development of double antibody sandwich ELISA

Objective To prepare the monoclonal antibody (McAb) against spike (S) protein of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and develop and verify a double antibody sandwich ELISA method for determination of spike protein antigen content. Methods McAb was prepared by hybridoma cell technology and identified. A double antibody sandwich ELISA using the prepared McAb as coating antibody and HRP-labeled rabbit anti-spike protein polyclonal antibody as detection antibody, and used for determination of SARS-CoV-2 spike protein antigen content. The concentrations of coating (10, 5, 2. 5 and 1. 25 microg / mL) and eniyme-labeled antibodies (4, 2 and l microg / mL) as well as kinds of bloc king reagents (non-blocked, 1% BSA diluted with PBS, 2% BSA diluted with PBS, 1% BSA + 1% sucrose, 2% BSA + 2% sucrose) were optimized. The developed method was verified .'or linear range, sensitivity, specificity and accuracy. The 10 samples with known S protein antigen contents at various stages of production process were determined by the developed method, of which the coincidence rate of result to that by a quantitative determination method developed by Institute of Medical Biology, Chinese Academy of Medical Sciences & Peking Union Medical College was calculated. Results Eighteen hybridom cell strains secreting S protein-specific McAb were screened. The ascites antibody prepared with 7B IOA2 cell line reached a concentration of 4. 487 mg/ mL and a purity of more than 90% after purification, which showed specific binding to the SI subunit of S protein. The McAb was of an antibody subtype of lgG2b, of which the titer and effect concentration were I: 128 000 and 0. 137 micro.g / mL respectively. The optimal concentrations of coating and enzyme-labeled antibodies were 5 and micro.g / mL respectively, while the optimal blocking reagent was 2% BSA + 2% sucrose. The linear range of the developed method was 2. 5 - 160 U, with a correlation coefficient (R2 ) of more than 0. 99 and a sensitivity of 1. 25 U. The method showed high specificity, with no reactions with nucleocapsid (N) protein, BSA, PBS and influenza virus. The recovery in verification for accuracy was 92. 10% - 111. 58%. The coincidence rates of determination results of samples with known S protein antigen contents by two methods were 94. 2% - 109% . Conclusion Specific McAb against S protein of SARS-CoV-2 was prepared, and a double antibody sandwich ELISA method was developed, which was suitable for determination of S protein content in vaccine products samples at various stages of production process and other samples. Copyright © 2022 Changchun Institute of Biological Products. All rights reserved.

Heavy pollution incidents in the context of "low emission" in Beijing-Tianjin-Hebei around the Chinese New Year of 2020

Persistent air pollution in the Beijing-Tianjin-Hebei region (BTH) has become an extremely complex challenge due to the combined effects of industrial structure, regional characteristics, weather and climate, and development. Although China's air pollution levels have reduced significantly since the Airborne Pollution Prevention and Control Action Plan and the Blue Sky Protection Campaign were implemented, the BTH remains a sensitive and vulnerable area. Such large decrease in primary pollution was mainly attributed to the substantial reductions in economic activities and associated emissions during the 2019 novel coronavirus (COVID-19) lockdown, i.e., around the Chinese New Year of 2020. Yet two consecutive severely polluting weather processes occurred in the BTH around the Chinese New Year of 2020, which have seeded doubt among the Chinese public and policymakers regarding the current scientific understanding of the mechanisms of haze pollution. The causes of formation and maintenance of pollution processes can differ significantly. The formation and maintenance of heavy pollution weather is caused by various factors, which is a complex process. Thus, it is crucial to distinguish the contribution of emissions and meteorological conditions on polluting weather, as well as distinguish the contribution of various meteorological factors on the formation and maintenance of polluting weather, for conducting effective attribution diagnostic analysis in actually environmental and meteorological impact assessment operation systems, especially in areas that are sensitive or vulnerable to air pollution. Identifying the specific meteorological conditions formed by polluting weather and establishing a comprehensive model for discriminating atmospheric dynamics and thermodynamics can provide a scientific basis for improving numerical models for air pollution potential forecasting in the future. Therefore, in this study, we focused on two consecutive severely polluting weather processes in BTH around the 2020 Chinese New Year (January to February 2020), as an ideal and unique field experiment for the prevention and control of current severe air pollution. We explored the reasons for the formation and maintenance of continuous severely polluting weather in the context of "continuous emissions reduction" and "relatively low social activity levels" from the perspective of the abnormal structure of the high-low atmospheric circulation system. Based on comprehensive diagnostic analyses, we quantified the relative contribution of each key meteorological factor to the continuous severely polluting weather in BTH by using the standardized multiple linear regression method. The results indicated that stable maintenance of low-level coastal high-pressure systems led to higher relative humidity at ground level compared with normal years and blocking systems, which are two key meteorological factors that induced persistent polluting weather in BTH. The abnormally stable blocking situation provided a special circulation background for the occurrence and maintenance of persistent heavy air pollution in BTH. The continuous and stable easterly and southerly water vapor transportation structure provided the BTH with more moisture than normal years, and it was conducive to increased moisture absorption by aerosols, especially under blocking. The "subsidence warming" effect of the high-level blocking high-pressure system facilitated the production of a "warm cover" structure in the middle of the troposphere. The presence of the anomalous warm cover structure in the troposphere facilitated the establishment of stable and high humidity weather, which was conducive to the accumulation of pollutants and continued air pollution. Dynamic systems (blocking systems) and water vapor transportation factors directly explained 46.8% of the meteorological causes of persistent heavy air polluting weather events around the 2020 Chinese New Year in BTH.

[The differential diagnosis of pulmonary infiltrates in cancer patients during the outbreak of the 2019 novel coronavirus disease].

Objective: To investigate the principles of differential diagnosis of pulmonary infiltrates in cancer patients during the outbreak of novel coronavirus (2019-nCoV) by analyzing one case of lymphoma who presented pulmonary ground-glass opacities (GGO) after courses of chemotherapy. Methods: Baseline demographics and clinicopathological data of eligible patients were retrieved from medical records. Information of clinical manifestations, history of epidemiology, lab tests and chest CT scan images of visiting patients from February 13 to February 28 were collected. Literatures about pulmonary infiltrates in cancer patients were searched from databases including PUBMED, EMBASE and CNKI. Results: Among the 139 cancer patients who underwent chest CT scans before chemotherapy, pulmonary infiltrates were identified in eight patients (5.8%), five of whom were characterized with GGOs in lungs. 2019-nCoV nuclear acid testing was performed in three patients and the results were negative. One case was a 66-year-old man who was diagnosed with non-Hodgkin lymphoma and underwent CHOP chemotherapy regimen. His chest CT scan image displayed multiple GGOs in lungs and the complete blood count showed decreased lymphocytes. This patient denied any contact with confirmed/suspected cases of 2019-nCoV infection, fever or other respiratory symptoms. Considering the negative result of nuclear acid testing, this patient was presumptively diagnosed with viral pneumonia and an experiential anti-infection treatment had been prescribed for him. Conclusions: The 2019 novel coronavirus disease (COVID-19) complicates the clinical scenario of pulmonary infiltrates in cancer patients. The epidemic history, clinical manifestation, CT scan image and lab test should be taken into combined consideration. The 2019-nCoV nuclear acid testing might be applied in more selected patients. Active anti-infection treatment and surveillance of patient condition should be initiated if infectious disease is considered.

Research on Academic Discourse Power of Scientific Collaboration on COVID-19 Pandemic

The sudden outbreak of COVID-19 pandemic at the beginning of 2020 poses a significant threat to the health and safety of people worldwide. Given the speed and scope of the COVID-19 pandemic, countries around the world have carried out scientific collaboration in the fight against the COVID-19 pandemic. This paper divides academic discourse power into academic discourse right and academic discourse impact. The number of published papers reflects the discourse right, and the number of cites reflects the academic discourse impact. The visualization analysis of research papers from 2019 to 2020 describes the worldwide scientific collaboration on COVID-19, and the academic discourse power of authors, institutions, and countries can be studied from the perspective of scientific collaboration. We analyze the scientific collaboration of 27,851 papers related to COVID-19 published during 2019-2020 from the perspectives of authors, institutions, and countries by using HistCite and VosViewer. Pearson correlation analysis is used to study the correlation between scientific collaboration, the number of published papers, and the number of cites. Furthermore, we find that scientific collaboration positively correlates with academic discourse right and academic discourse impact. Based on analysis of author collaboration, institutional collaboration and country collaboration, it was concluded that China has the highest total cites, reflecting its high academic discourse impact during the COVID-19 pandemic, and the USA has the highest number of international collaborators and the highest total number of published papers, reflecting its high discourse right during the COVID-19 pandemic. The number of cites and the number of published papers are significantly positively correlated with the number of collaborators in Pearson correlation of author collaboration, institutional collaboration and country collaboration.This study has presented the global collaboration on the research of COVID-19. We compared academic discourse right and academic discourse impact across different levels of authors, institutions, and countries, concluding that academic discourse right and academic discourse impact are significantly positively correlated with the number of collaborators. © 2022 ACM.

Development and Application of Global Health Events-Mental Stress Scale for Assessment of Medical Staff?s Acute Mental Stress Responses

Background: Medical workers have been increasingly involved in emergent public health events, which can lead to severe stress. However, no standardized, officially recognized, unified tool exists for mental distress measurement in medical workers who experienced the public health events. Purpose: In the present study, we propose the Global Health Events-Mental Stress Scale (GHE-MSS), as a revised version of the Impact of Event Scale-Revision (IES-R), for assessment of medical workers' acute mental stress responses within one month and their chronic mental stress responses within six months after major health events. Patients and methods: The IES-R was slightly modified, developed, and its reliability and validity were tested using the Delphi survey, primary survey with 115 participants, formal survey with 300 participants, and clinical evaluation with 566 participants. Results: Exploratory factor analysis and confirmatory factor analysis confirmed a promising validity of the scale. The values of Cronbach's alpha coefficient, the Spearman-Brown coefficient, and the retested Cronbach's alpha coefficient of the scale applied for the clinical evaluation were 0.88, 0.87, and 0.98, respectively, which confirmed a good internal consistency and stability. The results of the goodness-of-fit test indicated a good adaptation of the model. A correlation analysis was conducted to assess the correlation between the GHE-MSS and the PCL-C, which had a correlation coefficient of 0.68 (P < 0.01). Conclusion: GHE-MSS can be applied with a promising reliability and validity for the assessment of the acute mental stress response of medical workers experiencing public health events. This method can also be used for the screening of mental stress-associated disorders.

COVID-19 Detection Method Based on Attentional Mechanism and LSTM

Since 2020, the Novel Coronavirus virus, which can cause upper respiratory and lung infections and even kill in severe cases, has been ravaging the globe. Rapid diagnostic tests have become one of the main challenges due to the severe shortage of test kits. This article proposes a model combining Long short-term Memory (LSTM) and Convolutional Block Attention Module for detection of COVID-19 from chest X-ray images. In this article, chest X-ray images from the COVID-19 radiology standard data set in the Kaggle repository were used to extract features by MobileNet, VGG19, VGG16 and ResNet50. CBAM and LSTM were used for classifcation detection. The simulation results showed that the experimental results showed that VGG16–CBAM–LSTM combination was the best combination to detect and classify COVID-19 from chest X-ray images. The classification accuracy of VGG-16-CBAM-LSTM combination was 95.80% for COVID-19, pneumonia and normal. The sensitivity and specificity of the combination were 96.54% and 98.21%. The F1 score was 94.11%. The CNN model proposed in this article contributes to automated screening of COVID-19 patients and reduces the burden on the healthcare delivery framework. © 2022, Springer Nature Switzerland AG.

The IMU augment SLAM on Unmanned Vehicle for detection of protective measures in COVID-19

Detecting protective measures (e.g., masks, goggles and protective clothing) is a momentous step in the fight against COVID-19. The detection mode of unmanned devices based on Simultaneous localization and mapping (SLAM) and fusion technology is more efficient, economical and safe than the traditional manual detection. In this paper, a tightly-coupled nonlinear optimization approach is used to augment the visual feature extraction of SLAM by the gyroscope of the IMU to obtain a high-precision visual inertial system for joint position and pose estimation. Based on the VINS-Mono frame, first, an LSD algorithm based on a conditional selection strategy is proposed to extract line features efficiently. Then, we propose recovering missing point features from line features. Moreover, we propose a strategy to recover vanishing point features from line features, and add residuals to the SLAM cost function based on optimization, which optimizes point-line features in real time to promote the tracking and matching accuracy. Second, the wavelet threshold denoising method based on the 3σcriterion is used to carry out real-time online denoising for gyroscope to improve the output precision. Our WD-PL-VINS was measured on publicly available EuRoC datasets, TUM VI datasets and evaluated and validated in lab testing with a unmanned vehicle (UV) based on the NVIDIA Jetson-TX2 development board. The results show that our method’s APE and RPE on MH 03 easy sequences are improved by 69.28% and 97.66%, respectively, compared with VINS-Mono. IEEE

A new method for detecting COVID-19 combining deep learning with improved histogram equalization and median filter

The new coronavirus (COVID-2019) epidemic outbreak has devastating impacts on people's daily lives and public healthcare systems. The chest X-ray image is an effective tool for diagnosing new coronavirus diseases. This paper proposes a new method to identify the new coronavirus from chest X-ray images to assist radiologists in fast and accurate image reading. We first enhance the contrast of X-ray images by using adaptive histogram equalization and eliminating image noise by using a median filter. Then, the X-ray image is fed to a sophisticated deep neural network (FAC-DPN-SENet) proposed by us to train a classifier, which is used to classify an X-ray image as usual or COVID-2019 or other pneumonia. Applying our method to a standard dataset, we achieve a classification accuracy of 93%, which is significantly better performance than several other state-of-the-art models, such as ResNet and DenseNet. This shows that the proposed method can be used as an effective tool to detect COVID-2019. © 2021 IEEE.

PIXEL-LEVEL AND AFFINITY-LEVEL KNOWLEDGE DISTILLATION FOR UNSUPERVISED SEGMENTATION OF COVID-19 LESIONS

Automatic segmentation of COVID-19 lesions is essential for computer-aided diagnosis. However, this task remains challenging because widely-used supervised based methods require large-scale annotated data that is difficult to obtain. Although an unsupervised method based on anomaly detection has shown promising results in [1], its performance is relatively poor. We address this problem by proposing a pixel-level and affinity-level knowledge distillation method. It obtains a pre-trained teacher network with rich semantic knowledge of CT images by constructing and training an auto-encoder at first, and then trains a student network with the same architecture as the teacher by distilling the teacher's knowledge only from normal CT images, and finally localizes COVID-19 lesions using the feature discrepancy between the teacher and the student networks. Besides, except for the traditional pixel-level distillation, we design the affinity-level distillation that takes into account the pairwise relationship of features to fully distill effective knowledge. We evaluate this method by using three different COVID-19 datasets and the experimental results show that the segmentation performance is largely improved when it is compared with the other existing unsupervised anomaly detection methods. © 2022 IEEE