Stealthy Physical Masked Face Recognition Attack via Adversarial Style Optimization (preprint)

Deep neural networks (DNNs) have achieved state-of-the-art performance on face recognition (FR) tasks in the last decade. In real scenarios, the deployment of DNNs requires taking various face accessories into consideration, like glasses, hats, and masks. In the COVID-19 pandemic era, wearing face masks is one of the most effective ways to defend against the novel coronavirus. However, DNNs are known to be vulnerable to adversarial examples with a small but elaborated perturbation. Thus, a facial mask with adversarial perturbations may pose a great threat to the widely used deep learning-based FR models. In this paper, we consider a challenging adversarial setting: targeted attack against FR models. We propose a new stealthy physical masked FR attack via adversarial style optimization. Specifically, we train an adversarial style mask generator that hides adversarial perturbations inside style masks. Moreover, to ameliorate the phenomenon of sub-optimization with one fixed style, we propose to discover the optimal style given a target through style optimization in a continuous relaxation manner. We simultaneously optimize the generator and the style selection for generating strong and stealthy adversarial style masks. We evaluated the effectiveness and transferability of our proposed method via extensive white-box and black-box digital experiments. Furthermore, we also conducted physical attack experiments against local FR models and online platforms.

Food Sources May Affect the Symptom Rates of COVID-19, an Epidemiological Analysis Based on the Public Data in Gansu Province, China, During the Summer Epidemic Cycle in 2022.

According to the public data collected from the Health Commission of Gansu Province, China, regarding the COVID-19 pandemic during the summer epidemic cycle in 2022, the epidemiological analysis showed that the pandemic spread stability and the symptom rate (the number of confirmed cases divided by the sum of the number of asymptomatic cases and the number of confirmed cases) of COVID-19 were different among 3 main epidemic regions, Lanzhou, Linxia, and Gannan; both the symptom rate and the daily instantaneous symptom rate (daily number of confirmed cases divided by the sum of daily number of asymptomatic cases and daily number of confirmed cases) in Lanzhou were substantially higher than those in Linxia and Gannan. The difference in the food sources due to the high difference of the population ethnic composition in the 3 regions was probably the main driver for the difference of the symptom rates among the 3 regions. This work provides potential values for prevention and control of COVID-19 in different regions.

Milk Small Extracellular Vesicles-derived miRNAs Restrict Porcine Epidemic Diarrhea Virus Infectivity (preprint)

Background: Virus-caused diseases are a huge challenge to both animals and human beings, especially coronaviruses. Porcine epidemic diarrhea virus (PEDV), a coronavirus, causes acute diarrhea and up to 100% mortality in piglets less than three weeks of age. Maternal immunity provides protection for piglets in resisting PEDV infection. Small extracellular vesicles (sEV) contain bioactive molecules such as miRNAs to exchange genetic and epigenetic information between cells. Our previous study suggested that milk sEV facilitated intestinal tract development and prevented LPS-induced intestine damage. However, the effects of milk sEV on the inhibition of viral infections remain unclear. Results: In this study, through in vivo experiments, we found that porcine milk sEV protected piglets from PEDV-induced diarrhea and death. In vitro, we clarified that this protective effect was partly generated through the inhibition of the PEDV-N protein and HMGB1 by sEV miR-let-7e and miR-27b, respectively. Conclusions: In conclusion, we report that porcine milk sEVs protected piglets from PEDV-induced diarrhea and death by inhibiting virus replication, and this protective effect was partly generated through the inhibition of the PEDV-N and HMGB1 pathways by exosomal miR-let-7e and miR-27b. This study reveals a new antiviral function of milk sEVs, and the results suggest that milk sEVs may act as a mother-offspring transmission pathway for protecting newborns against PEDV infection.

N439K variant in spike protein may alter the infection efficiency and antigenicity of SARS-CoV-2 based on molecular dynamics simulation (preprint)

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), causing an outbreak of coronavirus disease 2019 (COVID-19), has been undergoing various mutations. The analysis of the structural and energetic effects of mutations on protein-protein interactions between the receptor binding domain (RBD) of SARS-CoV-2 and angiotensin converting enzyme 2 (ACE2) or neutralizing monoclonal antibodies will be beneficial for epidemic surveillance, diagnosis, and optimization of neutralizing agents. According to the molecular dynamics simulation, a key mutation N439K in the SARS-CoV-2 RBD region created a new salt bridge which resulted in greater electrostatic complementarity. Furthermore, the N439K-mutated RBD bound hACE2 with a higher affinity than wild-type, which may lead to more infectious. In addition, the N439K-mutated RBD was markedly resistant to the SARS-CoV-2 neutralizing antibody REGN10987, which may lead to the failure of neutralization. These findings would offer guidance on the development of neutralizing antibodies and the prevention of COVID-19.

An improved multivariate model that distinguishes COVID-19 from seasonal flu and other respiratory diseases.

COVID-19 shared many symptoms with seasonal flu, and community-acquired pneumonia (CAP) Since the responses to COVID-19 are dramatically different, this multicenter study aimed to develop and validate a multivariate model to accurately discriminate COVID-19 from influenza and CAP. Three independent cohorts from two hospitals (50 in discovery and internal validation sets, and 55 in the external validation cohorts) were included, and 12 variables such as symptoms, blood tests, first reverse transcription-polymerase chain reaction (RT-PCR) results, and chest CT images were collected. An integrated multi-feature model (RT-PCR, CT features, and blood lymphocyte percentage) established with random forest algorism showed the diagnostic accuracy of 92.0% (95% CI: 73.9 - 99.1) in the training set, and 96. 6% (95% CI: 79.6 - 99.9) in the internal validation cohort. The model also performed well in the external validation cohort with an area under the receiver operating characteristic curve of 0.93 (95% CI: 0.79 - 1.00), an F1 score of 0.80, and a Matthews correlation coefficient (MCC) of 0.76. In conclusion, the developed multivariate model based on machine learning techniques could be an efficient tool for COVID-19 screening in nonendemic regions with a high rate of influenza and CAP in the post-COVID-19 era.

An Innovative Non-Pharmaceutical Intervention to Mitigate SARS-CoV02 Spread: Probability Sampling to Identify and Isolate Asymptomatic Cases (preprint)

Studies estimate that a substantial proportion of SARS-CoV-2 transmission occurs through individuals who do not exhibit symptoms. Mitigation strategies test only those who are moderately to severely symptomatic, excluding the substantial portion of cases that are asymptomatic yet still infectious and likely responsible for a large proportion of the virus spread (1-8). While isolating asymptomatic cases will be necessary to effectively control viral spread, these cases are functionally invisible and there is no current method to identify them for isolation. To address this major omission in COVID-19 control, we develop a strategy, Sampling-Testing-Quarantine (STQ), for identifying and isolating individuals with asymptomatic SARS-CoV-2 in order to mitigate the epidemic. STQ uses probability sampling in the general population, regardless of symptoms, then isolates the individuals who test positive along with their household members who are high probability for asymptomatic infections. To test the potential efficacy of STQ, we use an agent-based model, designed to computationally simulate the epidemic in the Seattle with infection parameters, like R0 and asymptomatic fraction, derived from population data. Our results suggest that STQ can substantially slow and decrease the spread of COVID-19, even in the absence of school and work shutdowns. Results also recommend which sampling techniques, frequency of implementation, and population subject to isolation are most efficient in reducing spread with limited numbers of tests.

The strategies of perioperative management in orthopedic department during the pandemic of COVID-19 (preprint)

Background: Coronavirus disease 2019 (COVID-19) broken out and spread rapidly nationwide at the beginning of 2020, which has brought huge impacts to people and work. The current situation of prevention and control is severe and urge guidance for clinicians, especially for medical systems. In the hope of providing a reference and recommendation for the prevention and control of the COVID-19, we carried out research to improve the quality of patients care and prevention during this epidemic. Methods: : All of the staff were trained rapidly to master personal protection in our department. We reviewed the patients’ discharged records who underwent surgery in our department during January 1st to March 1st in 2019 and January 1st to March 1st in 2020. The managements of the surgery patients and flow charts were described and analyzed. Post-operation outcomes of the patients including duration, complications, surgical site infection (SSI), system infection, re-operation, and mortality. Both chi-squared test and Student’s t-test were performed to determine the relationship between the two periods in term of post-operation outcomes. Results: : Descriptive statistics analysis revealed that demographic of the patients between the two periods is similar. We had been benefited from the strict flow charts, smart robot and protection equipment in management of perioperative for orthopedic patients. With the help of the strict flow charts and smart equipment, post-operation outcomes of the patients revealed that the rates of the complications and re-operation had been reduced significantly ( p ＜0.05), while duration of operation, SSI and system infection had no significantly difference between two periods ( p ＞0.05). No patient and staff caught COVID-19 infection or mortality during the epidemic. Conclusions: : Our study indicated that medical quality and efficiency were affected little with the help of strategies described above during the epidemic, which could be a reference tool for medical staff in routine clinical practice for admission of patients around the world. What’s more, the provided strategies, which may evolve over time, could be used as empirical guidance and reference for orthopedic peers to get through the pandemic and ensure the normal operation of the hospital.

Perceptions and Behavioral Adoptions of the Preventive Strategies from Person-to-Person Transmission of COVID-19 Among the Public in China: An Online Cross-Sectional Survey (preprint)

Background: The strategies adopted to prevent spreading of COVID-19 are quarantine, social distancing and isolation of infected cases. This study investigates perceptions and behavioral adoptions of COVID-19 prevention strategies among the Chinese public and identified factors predicting individual health behavior.Methods: We conducted a cross-sectional online survey between 22 February and 5 March, 2020. We approached to urban residents aged over 18 years through snowball sampling method using the Chinese social media. The Health Belief Model was adopted to guide the analysis. Bivariate and multivariate logistic regressions were used to examine impacts of modifying factors (including demographic and socio-economic characteristics) and individual beliefs on individual health behavior.Results: of 5675 valid questionnaires, 95.8% of the respondents well understood the preventive measures from COVID-19 transmission, while 79.9% of the respondents adopted the behavior advised. 45.7% of the respondents perceived severity of the disease, 75.6% of the respondents perceived benefits of social constraints measures and 62.7% reported anxiety during the epidemic. After adjusting for modifying factors and individual beliefs, those who were female, had better income and good knowledge on preventive measures, perceived benefits on social constraint measures and did not feel anxiety were more likely to adopt behaviors advised.Conclusions: The Chinese public highly accepted and adopted behaviors advised to slow down the COVID-19 epidemic. People with low income or feeling anxiety were less likely to adopt the behavior advised. The policy support should target on the social vulnerable groups. The psychological support should be disseminated through different means, and the consultation should be provided to those who are in need.

Double-zero-event studies matter: a re-evaluation of physical distancing, face masks, and eye protection for preventing person-to-person transmission of COVID-19 and its policy impact (preprint)

ObjectivesHigh-quality meta-analyses on COVID-19 are in urgent demand for evidence-based decision making. However, conventional approaches exclude double-zero-event studies (DZS) from meta-analyses. We assessed whether including such studies impacts the conclusions in a recent systematic urgent review on prevention measures for preventing person-to-person transmission of COVID-19. Study designs and settingsWe extracted data for meta-analyses containing DZS from a recent review that assessed the effects of physical distancing, face masks, and eye protection for preventing person-to-person transmission. A bivariate generalized linear mixed model was used to re-do the meta-analyses with DZS included. We compared the synthesized relative risks (RRs) of the three prevention measures, their 95% confidence intervals (CI), and significance tests (at the level of 0.05) including and excluding DZS. ResultsThe re-analyzed COVID-19 data containing DZS involved a total of 1,784 participants who were not considered in the original review. Including DZS noticeably changed the synthesized RRs and 95% CIs of several interventions. For the meta-analysis of the effect of physical distancing, the RR of COVID-19 decreased from 0.15 (95% CI, 0.03 to 0.73) to 0.07 (95% CI, 0.01 to 0.98). For several meta-analyses, the statistical significance of the synthesized RR was changed. The RR of eye protection with a physical distance of 2 m and the RR of physical distancing when using N95 respirators were no longer statistically significant after including DZS. ConclusionsDZS may contain useful information. Sensitivity analyses that include DZS in meta-analysis are recommended.

Perceptions and Behavioral Adoptions of the Preventive Strategies from Person-to-Person Transmission of COVID-19 among the Public in China: An Online Cross-Sectional Survey (preprint)

Background: The strategies adopted to prevent spreading of COVID-19 are quarantine, social distancing and isolation of infected cases. This study investigates perceptions and behavioral adoptions of COVID-19 prevention strategies among the Chinese public and identified factors predicting individual health behavior.Methods: We conducted a cross-sectional online survey between 22 February and 5 March, 2020. We approached to urban residents aged over 18 years through snowball sampling method using the Chinese social media. The Health Belief Model was adopted to guide the analysis. Bivariate and multivariate logistic regressions were used to examine impacts of modifying factors (including demographic and socio-economic characteristics) and individual beliefs on individual health behavior.Results: of 5675 valid questionnaires, 95.8% of the respondents well understood the preventive measures from COVID-19 transmission, while 79.9% of the respondents adopted the behavior advised. 45.7% of the respondents perceived severity of the disease, 75.6% of the respondents perceived benefits of social constraints measures and 62.7% reported anxiety during the epidemic. After adjusting for modifying factors and individual beliefs, those who were female, had better income and good knowledge on preventive measures, perceived benefits on social constraint measures and did not feel anxiety were more likely to adopt behaviors advised.Conclusions: The Chinese public highly accepted and adopted behaviors advised to slow down the COVID-19 epidemic. People with low income or feeling anxiety were less likely to adopt the behavior advised. The policy support should target on the social vulnerable groups. The psychological support should be disseminated through different means, and the consultation should be provided to those who are in need.

The Relationship Between COVID-19 Infection and Risk Perception, Knowledge, Attitude As Well As Four Non-pharmaceutical Interventions (NPIs) During the Late Period Of The COVID-19 Epidemic In China An Online Cross-sectional Survey of 8158 Adults (preprint)

BackgroundSo far, there has been no published population study on the relationship between COVID-19 infection and publics risk perception, information source, knowledge, attitude and four non-pharmaceutical interventions(NPI: hand washing, proper coughing habits, social distancing and mask wearing) during the COVID-19 outbreak in China. MethodsAn online survey of 8158 Chinese adults between 22 February to 5 March 2020 was conducted. Bivariate associations between categorical variables were examined using Fisher exact test. We also explored the determinants of four NPIs as well as their association with COVID-19 infection using logistic regression. ResultsOf 8158 adults included, 57 (0.73%) were infected with COVID-19. The overwhelming majority of respondents showed a positive attitude (99.2%), positive risk perception (99.9%) and high knowledge levels that were among the strongest predictors of four highly adopted NPIs (hand washing:96.8%; proper coughing: 93.1%; social distancing:87.1%; mask wearing:97.9%). There was an increased risk of COVID-19 infection for those who not washing hands (2.28% vs 0.65%; RR=3.53: 95%CI: 1.53-8.15; P<0.009); not practicing proper coughing (1.79% vs 0.73%; RR=2.44: 95%CI: 1.15-5.15;P=0.026); not practicing social distancing (1.52% vs 0.58%; RR=2.63:95%CI:1.48 - 4.67; P=0.002); and not wearing a mask (7.41% vs 0.6%; RR=12.38:95%CI:5.81-26.36; P<0.001). For those who did practice all other three NPIs, wearing mask was associated with significantly reduced risk of infection compared to those who did not wear a mask (0.6% vs 16.7%; p=0.035). Similarly, for those who did not practice all or part of the other three NPIs, wearing mask was also associated with significantly reduced risk of infection. In a penalised logistic regression model including all four NPIs, wearing a mask was the only significant predictor of COVID-19 infection among four NPIs (OR=7.20; 95%CI:2.24-23.11; p<0.001). ConclusionsWe found high levels of risk perception, positive attitude, desirable knowledge as well as a high level of adopting four NPIs. The relevant knowledge, risk perception and attitude were strong predictors of adapting the four NPIs. Mask wearing, among four personal NPIs, was the most effective protective measure against COVID-19 infection with added preventive effect among those who practised all or part of the other three NPIs.