ABSTRACT
The pandemic outbreak of severe acute respiratory syndrome caused by the Coronavirus 2 disease in 2019, also known as SARS-COV-2 and COVID-19, has claimed over 5.6 million lives till now. The highly infectious nature of the Covid-19 virus has resulted into multiple massive upsurges in counts of new infections termed as 'waves.' These waves consist of numerous rising and falling counts of Covid-19 infection cases with changing dates that confuse analysts and researchers. Due to this confusion, the detection of emergence or drop of Covid waves is currently a subject of intensive research. Hence, we propose an algorithmic framework to forecast the upcoming details of Covid-19 infection waves for a region. The framework consists of a displaced double moving average ( δ DMA) algorithm for forecasting the start, rise, fall, and end of a Covid-19 wave. The forecast is generated by detection of potential dates with specific counts called 'markers.' This detection of markers is guided by decision rules generated through rough set theory. We also propose a novel 'corrected moving average' ( χ SMA) technique to forecast the upcoming count of new infections in a region. We implement our proposed framework on a database of Covid-19 infection specifics fetched from 12 countries, namely: Argentina, Colombia, New Zealand, Australia, Cuba, Jamaica, Belgium, Croatia, Libya, Kenya, Iran, and Myanmar. The database consists of day-wise time series of new and total infection counts from the date of first case till 31st January 2022 in each of the countries mentioned above. The δ DMA algorithm outperforms other baseline techniques in forecasting the rise and fall of Covid-19 waves with a forecast precision of 94.08%. The χ SMA algorithm also surpasses its counterparts in predicting the counts of new Covid-19 infections for the next day with the least mean absolute percentage error (MAPE) of 36.65%. Our proposed framework can be deployed to forecast the upcoming trends and counts of new Covid-19 infection cases under a minimum observation window of 7 days with high accuracy. With no perceptible impact of countermeasures on the pandemic until now, these forecasts will prove supportive to the administration and medical bodies in scaling and allotment of medical infrastructure and healthcare facilities.
ABSTRACT
Unlike prevalent facial expressions, micro expressions have subtle, involuntary muscle movements which are short-lived in nature. These minute muscle movements reflect true emotions of a person. Due to the short duration and low intensity, these micro-expressions are very difficult to perceive and interpret correctly. In this paper, we propose the dynamic representation of micro-expressions to preserve facial movement information of a video in a single frame. We also propose a Lateral Accretive Hybrid Network (LEARNet) to capture micro-level features of an expression in the facial region. The LEARNet refines the salient expression features in accretive manner by incorporating accretion layers (AL) in the network. The response of the AL holds the hybrid feature maps generated by prior laterally connected convolution layers. Moreover, LEARNet architecture incorporates the cross decoupled relationship between convolution layers which helps in preserving the tiny but influential facial muscle change information. The visual responses of the proposed LEARNet depict the effectiveness of the system by preserving both high- and micro-level edge features of facial expression. The effectiveness of the proposed LEARNet is evaluated on four benchmark datasets: CASME-I, CASME-II, CAS(ME)'2 and SMIC. The experimental results after investigation show a significant improvement of 4.03%, 1.90%, 1.79% and 2.82% as compared with ResNet on CASME-I, CASME-II, CAS(ME)'2 and SMIC datasets respectively.
