Masked Student Dataset of Expressions

Facial expression recognition (FER) algorithms work well in constrained environments with little or no occlusion of the face. However, real-world face occlusion is prevalent, most notably with the need to use a face mask in the current Covid-19 scenario. While there are works on the problem of occlusion in FER, little has been done before on the particular face mask scenario. Moreover, the few works in this area largely use synthetically created masked FER datasets. Motivated by these challenges posed by the pandemic to FER, we present a novel dataset, the Masked Student Dataset of Expressions or MSD-E, consisting of 1,960 real-world non-masked and masked facial expression images collected from 142 individuals. Along with the issue of obfuscated facial features, we illustrate how other subtler issues in masked FER are represented in our dataset. We then provide baseline results using ResNet-18, finding that its performance dips in the non-masked case when trained for FER in the presence of masks. To tackle this, we test two training paradigms: contrastive learning and knowledge distillation, and find that they increase the model's performance in the masked scenario while maintaining its non-masked performance. We further visualise our results using t-SNE plots and Grad-CAM, demonstrating that these paradigms capitalise on the limited features available in the masked scenario. Finally, we benchmark SOTA methods on MSD-E. The dataset is available at https://github.com/SridharSola/MSD-E. © 2022 ACM.

An Epidemiological Analysis of the Recipients of the First Dose of the First Phase of COVID-19 Vaccination in Oyo State, South-Western Nigeria.

BACKGROUND: Millions of people across the globe have been infected with coronavirus disease (COVID-19), and many lives have been lost in the process. As a result, vaccines are being developed to protect people from COVID-19 morbidity and mortality. Therefore, this study was conducted to assess the coverage rate for the COVID-19 vaccine in Oyo State. METHODS: A descriptive secondary analysis of COVID-19 immunization data was done between March and April 2021. Data were extracted from the original paper format and entered into Excel sheets. Charts and line graphs were plotted to determine the coverage rates. RESULTS: The overall coverage rate for the State was 81.0%. The highest and lowest coverage rates were 243.0% and 39.0% for Ibadan North and Iseyin Local Government Areas (LGAs), respectively. The proportion of female health workers vaccinated in the State was 64.5%. The proportion of male strategic leaders and male frontline workers was 62.5% and 55.7%, respectively. Akinyele and Egbeda LGAs recorded the same highest number of cases (27) of adverse events following immunization (AEFI). CONCLUSION: The study highlights the high proportion of vaccinated people in the State, while there was a low proportion of vaccinees in some LGAs. Therefore, effort to scale-up coverage across all the LGAs is recommended.

CONTEXTE: Des millions de personnes dans le monde ont été infectées par le COVID-19 et de nombreuses vies ont été perdues dans ce processus. En conséquence, des vaccins sont en cours de développement pour protéger les personnes contre la morbidité et la mortalité liées au COVID-19. Cette étude a donc été menée pour évaluer le taux de couverture du vaccin COVID-19 dans l'Etat d'Oyo. MÉTHODES: Une analyse secondaire descriptive des données de vaccination COVID-19 a été réalisée entre mars et avril 2021. Les données ont été extraites du format papier original et saisies dans des feuilles Excel. Des diagrammes et des graphiques linéaires ont été tracés pour déterminer les taux de couverture. RÉSULTATS: Le taux de couverture global de l'État était de 81,0 %. Les taux de couverture les plus élevés et les plus faibles étaient respectivement de 243,0 % et 39,0 % pour les zones de gouvernement local (LGA) d'Ibadan Nord et d'Iseyin. La proportion d'agents de santé féminins vaccinés dans l'État était de 64,5 %. La proportion d'hommes leaders stratégiques et d'hommes travailleurs de première ligne était respectivement de 62,5 % et 55,7 %. Les LGA d'Akinyele et d'Egbeda ont enregistré le même nombre élevé de cas (27) d'événements indésirables après la vaccination (EIAS). CONCLUSION: L'étude met en évidence la forte proportion de personnes vaccinées dans l'état, alors qu'il y avait une faible proportion de vaccinés dans certaines zone de gouvernement local. Il est donc recommandé de déployer des efforts pour augmenter la couverture vaccinale dans toutes les AGL. Mots clés: Épidémiologie, COVID-19, vaccin, première phase, Nigéria.

Priority in interhospital transfers of patients with severe COVID-19: development and prospective validation of a triage tool

OBJECTIVES: To develop and validate a triage scale (Spanish acronym, TIHCOVID) to assign priority by predicting critical events in patients with severe COVID-19 who are candidates for interhospital transfer. MATERIAL AND METHODS: Prospective cohort study in 2 periods for internal (February-April 2020) and external (October-December 2020) validation. We included consecutive patients with severe COVID-19 who were transported by the emergency medical service of Catalonia. A risk model was developed to predict mortality based on variables recorded on first contact between the regional emergency coordination center and the transferring hospital. The model's performance was evaluated by means of calibration and discrimination, and the results for the first and second periods were compared. RESULTS: Nine hundred patients were included, 450 in each period. In-hospital mortality was 33.8%. The 7 predictors included in the final model were age, comorbidity, need for prone positioning, renal insufficiency, use of high-flow nasal oxygen prior to mechanical ventilation, and a ratio of PaO2 to inspired oxygen fraction of less than 50. The performance of the model was good (Brier score, 0.172), and calibration and discrimination were consistent. We found no significant differences between the internal and external validation steps with respect to either the calibration slopes (0.92 [95% CI, 0.91-0.93] vs 1.12 [95% CI, 0.6-1.17], respectively;P = .150) or discrimination (area under the curve, 0.81 [95% CI, 0.75-0.84] vs 0.85 [95% CI, 0.81-0.89];P = .121). CONCLUSION: The TIHCOVID tool may be useful for triage when assigning priority for patients with severe COVID-19 who require transfer between hospitals.