ABSTRACT
With the worldwide spread of COVID-19, people are trying different ways to prevent the spread of the virus. One of the most useful and popular ways is wearing a face mask. Most people wear a face mask when they go out, which makes facial expression recognition become harder. Thus, how to improve the performance of the facial expression recognition model on masked faces is becoming an important issue. However, there is no public dataset that includes facial expressions with masks. Thus, we built two datasets which are a real-world masked facial expression database (VIP-DB) and a man-made masked facial expression database (M-RAF-DB). To reduce the influence of masks, we utilize contrastive representation learning and propose a two-branches network. We study the influence of contrastive learning on our two datasets. Results show that using contrastive representation learning improves the performance of expression recognition from masked face images. © 2022 ACM.
ABSTRACT
Purpose : Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has already been detected in ocular samples. However, the role of the eye in Coronavirus disease 2019 (COVID-19) is still unclear. We investigated the presence of SARS-CoV-2 in conjunctival swabs from patients with confirmed severe form of COVID-19 searching for differences in the presentation and assay positivity. Methods : This cross-sectional study included 50 conjunctival swab samples (one eye per patient) collected from 50 patients with confirmed COVID-19 in Hospital das Clínicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, Brazil. Samples were collected within 24 hours from the naso/oropharyngeal swab. Inclusion criteria was severe/critical disease and indication for hospitalization according to the Institutional Guideline, which was the presence of Severe Acute Respiratory Syndrome. Conjunctival swab was collected from the inferior fornix, without anesthesia, using a nylon flocked swab. Real-time reverse transcription polymerase chain reaction (rRT-PCR) was performed with primers and probes described in CDC protocol. The study was approved by the Institutional Ethics Committee and informed consent was obtained prior to procedures. Results : Twenty-four patients were male and twenty-six were female. Median age was 57.38 years (SD 15.23). Mean duration of symptoms before ocular sampling was 7.6 days (SD 3.52). Naso/oropharyngeal swab RT-PCR (within 24 hours from the conjunctival swab) was positive in 34 (68%) patients, negative in 14 (28%) and inconclusive in 2 (4%) patients. All the patients with negative or inconclusive RT-PCR had COVID-19 diagnosis confirmed by another naso/oropharyngeal swab or serology. Five (10%) conjunctival swabs resulted in positive rRT-PCR analysis and two (4%) had inconclusive results. Among the positive patients, 2 patients had a negative naso/oropharyngeal swab and 1 patient had an inconclusive result. None of the patients had ocular symptoms. Conclusions : The positivity rate of conjunctival swab for SARS-CoV-2 was higher when compared to previous studies and similar to those that evaluated moderate to severe forms of COVID-19, and was not associated with ocular symptoms. This suggests that a greater viral load may be present in the tear film of patients with severe form of COVID-19 and may not be associated with conjunctivitis. Special concern should be taken to healthcare workers.