LLDSNet: Limited labeled data segmentation network for COVID-19 CT lesion segmentation

With the COVID-19 outbreak in 2019, the world is facing a major crisis and people's health is at serious risk. Accurate segmentation of lesions in CT images can help doctors understand disease infections, prescribe the right medicine and control patients' conditions. Fast and accurate diagnosis not only can make the limited medical resources get reasonable allocation, but also can control the spread of disease, and computer-aided diagnosis can achieve this purpose, so this paper proposes a deep learning segmentation network LLDSNet based on a small amount of data, which is divided into two modules: contextual feature-aware module (CFAM) and shape edge detection module (SEDM). Due to the different morphology of lesions in different CT, lesions with dispersion, small lesion area and background area imbalance, lesion area and normal area boundary blurred, etc. The problem of lesion segmentation in COVID-19 poses a major challenge. The CFAM can effectively extract the overall and local features, and the SEDM can accurately find the edges of the lesion area to segment the lesions in this area. The hybrid loss function is used to avoid the class imbalance problem and improve the overall network performance. It is demonstrated that LLDSNet dice achieves 0.696 for a small number of data sets, and the best performance compared to five currently popular segmentation networks. © 2023 SPIE.

Analysis of the Spatial Distribution and Associated Factors of the Transmission Locations of COVID-19 in the First Four Waves in Hong Kong

Understanding the space–time pattern of the transmission locations of COVID-19, as well as the relationship between the pattern, socioeconomic status, and environmental factors, is important for pandemic prevention. Most existing research mainly analyzes the locations resided in or visited by COVID-19 cases, while few studies have been undertaken on the space–time pattern of the locations at which the transmissions took place and its associated influencing factors. To fill this gap, this study focuses on the space–time distribution patterns of COVID-19 transmission locations and the association between such patterns and urban factors. With Hong Kong as the study area, transmission chains of the four waves of COVID-19 outbreak in Hong Kong during the time period of January 2020 to June 2021 were reconstructed from the collected case information, and then the locations of COVID-19 transmission were inferred from the transmission chains. Statistically significant clusters of COVID-19 transmission locations at the level of tertiary planning units (TPUs) were detected and compared among different waves of COVID-19 outbreak. The high-risk areas and the associated influencing factors of different waves were also investigated. The results indicate that COVID-19 transmission began with the Hong Kong Island, further moved northward towards the New Territories, and finally shifted to the south Hong Kong Island, and the transmission population shows a difference between residential locations and non-residential locations. The research results can provide health authorities and policy-makers with useful information for pandemic prevention, as well as serve as a guide to the public in the avoidance of activities and places with a high risk of contagion. © 2023 by the authors.

A Novel Framework to Forecast COVID-19 Incidence Based on Google Trends Search Data

The global outbreak of coronavirus disease 2019 (COVID-19) has spread to more than 200 countries worldwide, leading to severe health and socioeconomic consequences. As such, the topic of monitoring and predicting epidemics has been attracting a lot of interest. Previous work reported search volumes from Google Trends are beneficial in decoding influenza dynamics, implying its potential for COVID-19 prediction. Therefore, a predictive model using the Wiener methods was built based on epidemic-related search queries from Google Trends, along with climate variables, aiming to forecast the dynamics of the weekly COVID-19 incidence in Washington, DC, USA. The Wiener model, which shares the merits of interpretability, low computation costs, and adaptation to nonlinear fluctuations, was used in this study. Models with multiple sets of features were constructed and further optimized by the highest weight selecting strategy. Furthermore, comparisons to the other two commonly used prediction models based on the autoregressive integrated moving average (ARIMA) and long short-term memory (LSTM) were also performed. Our results showed the predicted COVID-19 trends significantly correlated with the actual (rho <inline-formula> <tex-math notation="LaTeX">$=$</tex-math> </inline-formula> 0.88, <inline-formula> <tex-math notation="LaTeX">$p $</tex-math> </inline-formula> <inline-formula> <tex-math notation="LaTeX">$<$</tex-math> </inline-formula> 0.0001), outperforming those with ARIMA and LSTM approaches, indicating Google Trends data as a useful tool in terms of COVID-19 prediction. Also, the model using 20 search queries with the highest weighting outperformed all other models, supporting the highest weight feature selection as a feasible criterion. Google Trends search query data can be used to forecast the outbreak of COVID-19, which might assist health policymakers to allocate health care resources and taking preventive strategies. IEEE

Going viral: A systematic review exploring the impact of COVID- 19 on medical student education in urology

Introduction & Objectives: During the COVID-19 pandemic, medical students across the globe were temporarily dismissed from clinical placements. In the field of urology, prevocational exposure is closely linked to medical students' interest in pursuing urology as a career and their confidence in dealing with urological conditions post-graduation. This systemic review evaluates the emerging educational interventions to improve urological knowledge and urology exposure among medical students in the context of a global pandemic. Method(s): A modified PRISMA (Preferred Reporting Items for Systematic Review and Meta- Analyses) search was conducted using MEDLINE and EMBASE were searched with keywords related to urology education, medical education, and medical students. The inclusion criteria were all English language articles in peer-reviewed publications from January 2020- current. Full-text articles were retrieved, evaluated, and included in the final analysis. Result(s): In total, 1255 records were identified through the initial literature search, and 21 full-text articles were reviewed for eligibility. Eight studies met the selection criteria and were included in this review. Most studies were conducted in the United States. All studies utilised online learning platforms or videoconferencing applications as part of their interventions. All studies implemented a combination of interventions, including case-based learning, didactic lectures, and online discussion boards. All studies reported at least one positive finding on Kirkpatrick level 1 or 2. (See table 1 for complete data extraction). Conclusion(s): A wide variety of effective educational interventions has been implemented since 2020 to ensure adequate urology education for medical students. The pandemic largely drove the broad adoption of online learning, and these online resources should be incorporated into pre-existing Australian and New Zealand urology curricula post- COVID, given their effectiveness and popularity among medical students internationally. There was, however, a lack of educational outcomes assessed at higher Kirkpatrick levels. A robust methodology and a larger sample size are needed for future studies.

Prevalence and risk factors of erectile dysfunction in COVID-19 patients: a systematic review and meta-analysis.

PURPOSE: Studies have found that erectile dysfunction (ED) may be a short-term or long-term complication in coronavirus disease 2019 (COVID-19) patients, but no relevant studies have completed a pooled analysis of this claim. The purpose of the review was to comprehensively search the relevant literature, summarize the prevalence of ED in COVID-19 patients, assess risk factors for its development, and explore the effect of the COVID-19 infection on erectile function. METHODS: Medline, Embase, and the Cochrane Library was performed from database inception until April 14, 2022. Heterogeneity was analyzed by χ2 tests and I2 was used as a quantitative test of heterogeneity. Subgroup analyses, meta-regression, and sensitivity analyses were used to analyze sources of heterogeneity. RESULTS: Our review included 8 studies, 4 of which functioned as a control group. There were 250,606 COVID-19 patients (mean age: 31-47.1 years, sample size: 23-246,990). The control group consisted of 10,844,200 individuals (mean age: 32.76-42.4 years, sample size 75-10,836,663). The prevalence of ED was 33% (95% CI 18-47%, I2 = 99.48%) in COVID-19 patients. The prevalence of ED based on the international coding of diseases (ICD-10) was 9% (95% CI 2-19%), which was significantly lower than the prevalence of ED diagnosed based on the International Index of Erectile Function (IIEF-5) (46%, 95% CI 22-71%, I2 = 96.72%). The pooling prevalence of ED was 50% (95% CI 34-67%, I2 = 81.54%) for articles published in 2021, significantly higher than that for articles published in 2022 (17%, 95% CI 7-30%, I2 = 99.55%). The relative risk of developing ED was 2.64 times in COVID-19 patients higher than in non-COVID-19 patients (RR: 2.64, 95% CI 1.01-6.88). The GRADE-pro score showed that the mean incidence of ED events in COVID-19 patients was 1,333/50,606 (2.6%) compared with 52,937/844,200 (0.4%) in controls; the absolute impact of COVID-19 on ED was 656/100,000 (ranging from 4/100,000 to 2352/100,000). Anxiety (OR: 1.13, 95% CI 1.03-1.26, I2 = 0.0%) in COVID-19 patients was a risk factor for ED. CONCLUSION: COVID-19 patients have a high risk and prevalence of ED, mainly driven by anxiety. Attention should be paid to patient's erectile functioning when treating COVID-19.

COVID-19 Impacts on Border Community Organizations

The COVID-19 pandemic massively affected the nonprofit sector. This article explores how the crisis has impacted nonprofit organizations at a U.S.-Mexico border community with a large population of minorities and migrants. Guided by resource dependency theory and the nonprofit capacity building framework, surveys reveal that nonprofits with less financial support from the government sector, low leadership, and weak operational capacities receive critical impacts from the pandemic. The findings also show that local nonprofits are bonded closely to the community during the pandemic, which reflects the collectivistic culture in Hispanic/Latino communities. This study provides important insights on how local nonprofits with limited resources and an increase in demand from vulnerable populations struggled with the pandemic.

Impact of COVID-19 Pandemic on the Frequency of Acute Exacerbations of Chronic Obstructive Pulmonary Disease

Background Respiratory virus infection is an important trigger of acute exacerbation of chronic obstructive pulmonary disease（AECOPD）. China has adopted a series of containment measures assisting to curb COVID-19 transmission since the outbreak of the pandemic. Several studies showed a decrease in hospitalizations for AECOPD during the COVID-19 pandemic. However，there has been a relative lack of studies investigating the effects of preventive measures on the frequency and severity of exacerbations. Objective To explore the impact of the COVID-19 pandemic on the frequency of AECOPD with or without medical attention. Methods The subjects were from a prospective COPD cohort study conducted in the First Affiliated Hospital of Guangzhou Medical University，which began recruiting patients in early 2016，with visits every 3 months to collect demographic and clinical data，including those who were followed up during June to August 2017（group 1），June to August 2018（group 2），June to August 2019 （group 3），and June to August 2020（group 4）. Basic clinical data （including the frequency of AECOPD，sex，age，symptom score and so on） were collected from group 1 from October 2016 to May 2017，group 2 from October 2017 to May 2018，group 3 from October 2018 to May 2019，and group 4 from October 2019 to May 2020（during which the periods from October 2019 to January 2020，and from February to May 2020 were defined as preCOVID-19 period，and post-COVID-19 period，respectively）. The frequency of AECOPD during October to May next year in group 4 was compared with that of the other three groups. The changes in the frequency of AECOPD between pre- and postCOVID-19 periods were analyzed. Results There were 162 patients in group 1，157 in group 2，167 in group 3，and 159 in group 4. Group 1 had a higher frequency of AECOPD in February to May than in October to January next year（P=0.013），so did group 2（P=0.016）. In contrast，group 4 had a higher frequency of AECOPD in October to January next year than in February to May（P=0.001）. The frequency of AECOPD during October to December in group 4 was similar to that of the other three groups（P>0.05）. But the frequency of AECOPD from February to April in group 4 was lower than that in groups 1-3 （P<0.05）. There was no significant difference in the monthly frequency of AECOPD without medical attention in group 4 compared with that of groups 1-3（P>0.05）. The frequency of AECOPD with medical attention from October to December in group 4 was similar to that of groups 1-3（P>0.05）. but it from February to April in group 4 was lower than that in groups 1-3（P<0.05）. Conclusion Prevention and control measures targeting COVID-19 may be contributive to reducing the frequency of AECOPD. It is suggested that COPD patients should reduce gathering activities，maintain social distance，wear masks when going out，and wash hands frequently even after the COVID-19. © 2023 Chinese General Practice. All rights reserved.

Identifying Risks in the Cruise Supply Chain: An Empirical Study in Shanghai, China

The increased frequency and severe consequences of risks in the cruise industry have attracted increasing attention from both academics and practitioners, especially after the 2012 ‘Costa Concordia' disaster and the 2020 coronavirus outbreak on the ‘Diamond Princess'. Although the literature on risk studies associated with the cruise industry and supply-chain risk management is growing, the extant literature lacks a study to view risks in the cruise industry associated with the supply chain. This paper addresses this gap by reviewing the literature on risks related to the cruise industry and general supply-chain risks to create a framework of cruise supply-chain risks. Then, semi-structured interviews were conducted to validate the identified risks and explore potential undiscovered risks. A novel risk typology of the cruise supply chain was then built based on the literature review and the empirical study. This includes macro risks, safety, security, and health risks, information risks, and supply risks. This framework can be applied for the purpose of systematically identifying the risks and their impacts on the cruise supply chain. This paper contributes to the development of a comprehensive cruise supply-chain risk classification with a detailed explanation of each risk in the cruise supply chain, which can be used by stakeholders in the cruise industry to identify and measure the impact of each risk. Additionally, this paper provides avenues for future research by scholars interested in assessing and managing cruise supply-chain risks. © National Academy of Sciences: Transportation Research Board 2022.

Impact of COVID-19 Pandemic on the Frequency of Acute Exacerbations of Chronic Obstructive Pulmonary Disease

Background Respiratory virus infection is an important trigger of acute exacerbation of chronic obstructive pulmonary disease（AECOPD）. China has adopted a series of containment measures assisting to curb COVID-19 transmission since the outbreak of the pandemic. Several studies showed a decrease in hospitalizations for AECOPD during the COVID-19 pandemic. However，there has been a relative lack of studies investigating the effects of preventive measures on the frequency and severity of exacerbations. Objective To explore the impact of the COVID-19 pandemic on the frequency of AECOPD with or without medical attention. Methods The subjects were from a prospective COPD cohort study conducted in the First Affiliated Hospital of Guangzhou Medical University，which began recruiting patients in early 2016，with visits every 3 months to collect demographic and clinical data，including those who were followed up during June to August 2017（group 1），June to August 2018（group 2），June to August 2019 （group 3），and June to August 2020（group 4）. Basic clinical data （including the frequency of AECOPD，sex，age，symptom score and so on） were collected from group 1 from October 2016 to May 2017，group 2 from October 2017 to May 2018，group 3 from October 2018 to May 2019，and group 4 from October 2019 to May 2020（during which the periods from October 2019 to January 2020，and from February to May 2020 were defined as preCOVID-19 period，and post-COVID-19 period，respectively）. The frequency of AECOPD during October to May next year in group 4 was compared with that of the other three groups. The changes in the frequency of AECOPD between pre- and postCOVID-19 periods were analyzed. Results There were 162 patients in group 1，157 in group 2，167 in group 3，and 159 in group 4. Group 1 had a higher frequency of AECOPD in February to May than in October to January next year（P=0.013），so did group 2（P=0.016）. In contrast，group 4 had a higher frequency of AECOPD in October to January next year than in February to May（P=0.001）. The frequency of AECOPD during October to December in group 4 was similar to that of the other three groups（P>0.05）. But the frequency of AECOPD from February to April in group 4 was lower than that in groups 1-3 （P<0.05）. There was no significant difference in the monthly frequency of AECOPD without medical attention in group 4 compared with that of groups 1-3（P>0.05）. The frequency of AECOPD with medical attention from October to December in group 4 was similar to that of groups 1-3（P>0.05）. but it from February to April in group 4 was lower than that in groups 1-3（P<0.05）. Conclusion Prevention and control measures targeting COVID-19 may be contributive to reducing the frequency of AECOPD. It is suggested that COPD patients should reduce gathering activities，maintain social distance，wear masks when going out，and wash hands frequently even after the COVID-19. © 2023 Chinese General Practice. All rights reserved.

Development of Machine Learning Regression Model for COVID-19 Drug Target Prediction

There is a perennial need to identify novel, effective therapeutic agents to combat rising infections. Recently, prediction of therapeutic targets to decrease the impact of COVID-19 has posed an urgent challenge requiring innovative solutions. Successful identification of novel drug-target combinations may greatly facilitate drug development. To meet this need, we developed a COVID-19 drug target prediction model using machine learning approaches to quickly identify drug candidates for 18 COVID-19 protein targets. Specifically, we analyzed the performance of three prediction models to predict drug-target docking scores, which represents the strength of interactions between ligands and proteins. Docking scores were predicted for 300,457 molecules on 18 different COVID-19 related protein docking targets. Our proposed approach achieved a competitive performance with mathrm{R}-{2}=0.69,MAE=0.285, MSE=0.627. In addition, we identify chemical structures associated with stronger binding affinities across target binding sites. We believe our work could potentially save pharmaceutical companies significant resources, especially during the early stages of drug development. © 2022 IEEE.

Using machine learning to understand Twitter users' urban green space activities during COVID-19 pandemic period

Volunteered Geographic Information (VGI) provides effective information for evaluating the usage of urban green space (UGS). Geo-referenced Tweets become very popular in the assessment of UGS use because of data availability and large data volume compared with traditional surveying methods, which are time-consuming and inefficient. However, previous studies lack efficient methods to extract and interpret Twitter data for UGS activities evaluation. Therefore, this paper aims to present a framework that enables high-efficient extraction of public UGS activities from Twitter. Greater London was selected as a case study to describe the framework development. First, Twitter data within Greater London over a certain COVID-19 lockdown period are collected, cleaned and pre-processed. Second, word vector representations were generated using Word2vec model, and then document vector representations were obtained by using Doc2vec model. Next, all the Tweets were clustered by using K-means algorithm to reveal the UGS activities during lockdown period. The framework can be used as a tool for UGS planners and managers to enable a holistic understanding of public activities engagement in UGS and increase the degree of public participation in UGS management. © 2022 IEEE.

Attention-based Automated Chest CT Image Segmentation Method of COVID-19 Lung Infection

According to the World Health Organization, Artificial Intelligence (AI) technology may assist in COVID-19 management. However, existing image segmentation using AI suffers from a lack of accuracy and explainability, which prevents its adoption in actual clinical practice. In this paper, we investigated an attention-based image segmentation method for COVID-19 CT imaging with enhanced interpretation capabilities. Specifically, we developed U-Net architecture-based for segmentation with attention coefficients to produce a salient feature map. We use the DICE score and accuracy to perform a comprehensive model evaluation. We compared to other well-known methods such as Light U-Net, COPLE-Net, and Res U-Net and demonstrated that attention U-Net is superior for COVID-19 segmentation tasks in terms of performance and explainability. We also developed the tool as a web-application with a graphic user interface with the goal to translate this AI-driven clinical decision-support system for real-world clinical use. © 2022 IEEE.

Hospital logistics management in the context of the COVID-19 epidemic -A Case Study of The First Affiliated Hospital of Xi'an Jiaotong University

Novel coronavirus pneumonia (NCP), known as COVID-19 for short, is a new infectious disease that spreads rapidly. Since the outbreak of the epidemic, the global epidemic situation is still grim. It is necessary for a comprehensive grade triple-A hospital to strictly control the transmission risk in the hospital and ensure the life and health safety of medical staff, other patients and accompanying staff, in addition to fulfilling its responsibility of treating COVID-19 patients. In this context, unprecedented challenges have been posed to the routine work flow of the hospital, including personnel management, system construction, supply and logistics support, and infection control. The paper summarizes the experience in hospital logistics support management, aiming at exploring strategies for coping with the COVID-19 epidemic. Copyright © 2022, Editorial Board of Journal of Xi'an Jiaotong University (Medical Sciences). All right reserved.

Identification of covid-19 severity and associated genetic biomarkers based on scrna-SEQ data

Bio-marker identification for COVID-19 remains a vital research area to improve current and future pandemic responses. Innovative artificial intelligence and machine learning-based systems may leverage the large quantity and complexity of single cell sequencing data to quickly identify disease with high sensitivity. In this study, we developed a novel approach to classify patient COVID-19 infection severity using single-cell sequencing data derived from patient BronchoAlveolar Lavage Fluid (BALF) samples. We also identified key genetic biomarkers associated with COVID-19 infection severity. Feature importance scores from high performing COVID-19 classifiers were used to identify a set of novel genetic biomarkers that are predictive of COVID-19 infection severity. Treatment development and pandemic reaction may be greatly improved using our novel big-data approach. Our implementation is available on https://github.com/aekanshgoel/COVID-19-scRNAseq. © 2022 Owner/Author.

SARS-CoV-2 VARIANTS INCREASE KINETIC STABILITY of OPEN SPIKE CONFORMATIONS

Background: SARS-CoV-2 variants of concern harbor mutations in the Spike (S) glycoprotein that confer more efficient transmission and dampen the efficacy of COVID-19 vaccines and antibody therapies. S mediates virus entry and is the primary target for antibody responses, with structural studies of soluble S variants revealing an increased propensity towards conformations accessible to the human Angiotensin-Converting Enzyme 2 (hACE2) receptor. However, real-time observations of conformational dynamics that govern the structural equilibriums of the S variants have been lacking. Methods: Here, we report single-molecule Förster Resonance Energy Transfer (smFRET) studies of S variants of concern containing critical mutations, including D614G and E484K, in the context of virus particles. Results: Investigated variants were shown by smFRET to predominantly occupy more open hACE2-accessible conformations, agreeing with predictions from structures of soluble trimers. Additionally, S variants exhibited decelerated transitions from hACE2-accessible/bound states. Conclusion: Here, we provide the real-time dimension to distinct structures of Spikes in the context of virus particles and present the first experimental evidence of increased stability of Spike variants. Our finding of increased S kinetic stability in the open conformation provides a new perspective on SARS-CoV-2 adaptation to the human population.

Learning from heterogeneous data via contrastive learning: An application in multi-source covid-19 radiography

The ongoing COVID-19 pandemic has overloaded current healthcare systems, including radiology systems and departments. Machine learning-based medical imaging diagnostic approaches play an important role in tracking the spread of this virus, identifying high-risk patients, and controlling infections in real-time. Researchers aggregate radiographic samples from different data sources to establish a multi-source learning scheme to mitigate the insufficiency of COVID-19 samples from individual hospitals, especially in the early stage of the disease. However, data heterogeneity across different clinical centers with various imaging conditions is considered a significant limitation in model performance. This paper proposes a contrastive learning scheme for the automatic diagnosis of COVID-19 to effectively mitigate data heterogeneity in multi-source data and learn a robust and generalizable model. Inspired by advances in domain adaptation, we employ contrastive training objectives to promote intra-class cohesion across different data sources and inter-class separation of infected and non-infected cases. Extensive experiments on two public COVID-19 CT datasets demonstrate the effectiveness of the proposed method for tackling data heterogeneity problems with boosted diagnosis performance. Moreover, benefiting from the contrastive learning framework, our method can be generalized to solve data heterogeneity problems under a broader multi-source learning setting. © 2021 IEEE

Automated risk assessment of covid-19 patients at diagnosis using electronic healthcare records

COVID-19 causes significant morbidity and mortality and early intervention is key to minimizing deadly complications. Available treatments, such as monoclonal antibody therapy, may limit complications, but only when given soon after symptom onset. Unfortunately, these treatments are often expensive, in limited supply, require administration within a hospital setting, and should be given before the onset of severe symptoms. These challenges have created the need for early triage of patients likely to develop life-threatening complications. To meet this need, we developed an automated patient risk assessment model using a real-world hospital system dataset with over 17,000 COVID-positive patients. Specifically, for each COVID-positive patient, we generate a separate risk score for each of four clinical outcomes including death within 30 days, mechanical ventilator use, ICU admission, and any catastrophic event (a superset of dangerous outcomes). We hypothesized that a deep learning binary classification approach can generate these four risk scores from electronic healthcare records data at the time of diagnosis. Our approach achieves significant performance on the four tasks with an area under receiver operating curve (AUROC) for any catastrophic outcome, death within 30 days, ventilator use, and ICU admission of 86.7%, 88.2%, 86.2%, and 87.8%, respectively. In addition, we visualize the sensitivity and specificity of these risk scores to allow clinicians to customize their usage within different clinical outcomes. We believe this work fulfills a clear clinical need for early detection of objective clinical outcomes and can be used for early screening for treatment intervention. © 2021 IEEE

CMC-COV19D: Contrastive Mixup Classification for COVID-19 Diagnosis

Deep learning methods have been extensively investigated for rapid and precise computer-aided diagnosis during the outbreak of the COVID-19 epidemic. However, there are still remaining issues to be addressed, such as distinguishing COVID-19 in the complex scenario of multi-type pneumonia classification. In this paper, we aim to boost the COVID-19 diagnostic performance with more discriminative deep representations of COVID and non-COVID categories. We propose a novel COVID-19 diagnosis approach with contrastive representation learning to effectively capture the intra-class similarity and inter-class difference. Besides, we design an adaptive joint training strategy to integrate the classification loss, mixup loss, and contrastive loss. Through the joint loss function, we obtain the high-level representations which are highly discriminative in COVID-19 screening. Extensive experiments on two chest CT image datasets, i.e., CC-CCII dataset and COV19-CT-DB database, demonstrate the effectiveness of our proposed approach in COVID-19 diagnosis. Our method won the first prize in the ICCV 2021 Covid-19 Diagnosis Competition of AI-enabled Medical Image Analysis Workshop. Our code is publicly available at https://github.com/houjunlin/Team-FDVTS-COVID-Solution.

SafeCampus: Multimodal-Based Campus-Wide Pandemic Forecasting

The motivation of this work is to build a multimodal-based COVID-19 pandemic forecasting platform for a large-scale academic institution to minimize the impact of COVID-19 after resuming academic activities. The design of this multimodality work is steered by video, audio, and tweets. Before conducting COVID-19 prediction, we first trained diverse models, including traditional machine learning models (e.g., Naive Bayes, support vector machine, and TF-IDF) and deep learning models [e.g., long short-term memory (LSTM), MobileNetV2, and SSD], to extract meaningful information from video, audio, and tweets by 1) detecting and counting face masks, 2) detecting and counting cough for potential infected cases, and 3) conducting sentiment analysis based on COVID-19-related tweets. Finally, we fed the multimodal analysis results together with daily confirmed cases data and social distancing metrics into the LSTM model to predict the daily increase rate of confirmed cases for the next week. Important observations with supporting evidence are presented. © 1997-2012 IEEE.