Emotion Mining from Speech in the Covid-19 Era: An Exploratory Study

Speech is the most effective form of communication because it is not limited to just the linguistic components but carries the speaker's emotions laced within the components like tone of voice and cues like cries and sighs. This paper aims at studying the research done in the past and applying it to the Covid-19 era.The pandemic is of a great magnitude, affecting every aspect of life including emotions. This time period requires research in determining the most dominant emotions in conversations, to serve as a reference for future research and as a contrast to the research done in the past. Previous papers have identified emotions like happiness, anger, fear and sadness using feature extraction algorithms like MFCC (Mel Frequency Cepstral Coefficients and numerous classification algorithms like GMM (Gaussian Mixture Model), SVM (Support Vector Machine), KNN (K-Nearest-neighbor) and HMM (Hidden Markov Model). Some research has pointed towards ASR (Automatic Speech Recognition), N-Grams and vector space modeling. This paper aims at recognizing the most suitable algorithms for determining the pandemic specific emotions in speech. © 2021 IEEE.

Association of Chest CT Severity Scores and Vaccination Status in COVID-19 Disease: A Cross-sectional Study

Introduction: Coronavirus Disease 2019 (COVID-19) has emerged as a pandemic with substantial morbidity and mortality. While global efforts towards mitigating the infection are focused on the vaccination of population, studies are warranted to prove the efficacy of vaccine in prevention of infection or reducing the severity of infection in affected patients. The 25-point High Resolution Computed Tomography (HRCT) severity score has proved to be an effective tool in estimating the severity of lung infection and correlates with laboratory parameters and disease outcome. The HRCT scores hence provide an objective evidence to prove the efficacy of vaccines in vaccinated individuals by assessing the extent of lung involvement. Aim: To compare the chest CT severity score in vaccinated and unvaccinated COVID-19 infected patients. Materials and Methods: This cross-sectional study was conducted in the Department of Radiology, at Sri Jayadeva Institute of Cardiovascular Sciences and Research, Mysuru, Karnataka, India. The data of HRCT scores and vaccination status was collected during the month of April 2021 from patients who were suspected to have COVID-19 infection and underwent a chest HRCT scan. The severity of lung infection in vaccinated and unvaccinated individuals was compared based on the HRCT scores and the association between these variables were analysed. The association between the respective variables were studied using Fischer's-exact and Kruskal-Wallis tests. Results: The study involved a total of 178 subjects (males were 98), where 127 (71.3%) were unvaccinated and 51 (28.6%) were vaccinated with one or both doses {Covaxin (Bharat Biotech) vaccine or Covishield (Oxford-AstraZeneca) vaccine approved by Emergency Use Authorisation (EUA)}. The frequency of disease was least in 14 (7.9%) among fully vaccinated subjects. Severe COVID-19 associated pneumonia with severity score of 18 or more was seen in 7% of unvaccinated individuals, while none of the partial/fully vaccinated individuals had severe disease. The median CT severity score was significantly higher among unvaccinated patients compared to partially and fully vaccinated patients (p-value=0.001). Fully vaccinated patients had almost low CT severity score indicating mild form of disease. Conclusion: To the best of authors knowledge, this study is the first to describe the chest CT severity scores of vaccinated individuals in comparison with the unvaccinated COVID-19 infected patients. The disease severity was significantly higher among unvaccinated patients compared to partially and fully vaccinated patients. The present study has provided substantial evidence of vaccine efficacy in reducing the disease severity in COVID-19 infected patients.