Face Recognition at varying angles from distant CCTV Footage using Siamese Architecture

Face Recognition is a deeply studied and researched domain. There are quite large solutions and model architectures to tackle majority of the face recognition related concerns. In this work we come up with a more specific version of face recognition which can mainly be used to achieve long distance and natural limitations of CCTV identification in real world scenarios using existing methods to better modifications. The solution which paper proposes using deep learning can be used to recognise a person even if they wear face masks due to the Covid-19 pandemic. One-shot learning is incorporated which can be used to train the specific model with just one image per person of the individual to be recognized. The designed model is modified from Siamese network architecture trained in triplet loss function to achieve these requirements. © 2022 IEEE.

Impact of the COVID-19 pandemic on cervical cancer diagnosis and treatment in Botswana

Background: The COVID-19 pandemic rapidly altered cancer care delivery globally, providing a compelling opportunity to empirically study how these changes impacted persistent disparities in care. Cervical cancer is one of the most common female cancers worldwide, with approximately 90% of cases and deaths occurring in low- and middle-income countries (LMICs). In Botswana, a LMIC with a particularly high prevalence of HIV and cervical cancer, delays in cervical cancer diagnosis and treatment have been documented but is unknown how these delays may have been mitigated or exacerbated since the pandemic. Methods: The objective of this analysis is to evaluate patterns of cervical cancer diagnosis and treatment initiation before (January 2015-March 2020) and during the pandemic (April 2020-July 2021) using longitudinal clinical and patient-reported data from a cohort of over 1,000 patients receiving care for gynecologic cancers in Botswana. The primary outcome is timeliness of treatment defined by the number of days between first clinical visit and initiation of first-line treatment and categorized dichotomously (> 30 days classified as delay). Primary exposure is the time period (prepandemic and pandemic) defined by the month of first visit. We calculated unadjusted proportion of delays and covariates stratified by time period and used bivariate analysis to examine factors associated with each time period. We used multivariable logistic regression models to examine the association between delay and time period, adjusting for all covariates (age, stage, HIV status, rurality, screening history, and partner status). Results are presented as unadjusted proportions, adjusted odds ratios (AOR), and 95% confidence intervals. Results: Of the 1,200 patients treated for cervical cancer at the multidisciplinary clinic, 990 (82.5%) were diagnosed pre-pandemic and 210 (17.5%) during the pandemic. Among all patients with gynecologic cancers (n = 1,568), the proportion of patients with cervical cancer significantly decreased from 78.6% pre-pandemic to 68.0% during the pandemic (p < 0.001). In comparison to pre-pandemic, patients with cervical cancer during the pandemic were significantly less likely to have attended a screening clinic prior to their treatment (57.6% vs 15.3%;p < 0.001) and significantly more likely to experience treatment delays (61.6% vs 92.9%;p < 0.001). In the multivariable model, patients diagnosed during the pandemic had a 7-fold higher likelihood of treatment delays than those patients diagnosed pre-pandemic (AOR: 7.95;95% CI: 4.45-14.19). Conclusions: The pandemic significantly increased delays in treatment for nearly all patients with cervical cancer in Botswana. Given persistent global disparities in cervical cancer, there is a great need to implement evidence-based strategies for improving screening and timeliness of care in Botswana and other LMICs.

Predicting Covid-19 Positive Cases and Analysis on the Relevance of Features using SHAP (SHapley Additive exPlanation)

COVID-19 and related infections are on the surge around the world, posing new threats to our society. There is a clear motivation to implement protective measures that aid in the effective control of future outbreaks or pandemics. The effect of the COVID-19 pandemic has prompted a flood of studies aimed at deeper understanding, monitoring, and also controlling the disease. Machine learning is increasingly becoming more prevalent in the area of medical diagnosis. With this paper, we will classify whether the patient is affected with covid or not and elucidate the significance of every attribute on the output using SHAP (SHapley Additive exPlanation). © 2021 IEEE.