Bebortta Sujit, Dalabehera Aditya Ranjan, Pati Bibudhendu, Panigrahi Chhabi Rani, Nanda Gyana Ranjan, Sahu Biswajit, Senapati Dilip
Department of Computer Science, Ravenshaw University, Cuttack, 753003, Odisha, India.
Department of Computer Science, Rama Devi Women's University, Bhubaneswar, 751022, Odisha, India.
Smart Health (Amst). 2022 Dec;26:100308. doi: 10.1016/j.smhl.2022.100308. Epub 2022 Aug 12.
In recent times, several strategies to minimize the spread of 2019 novel coronavirus disease (COVID-19) have been adopted. Some recent technological breakthroughs like the drone-based tracking systems have focused on devising specific dynamical approaches for administering public mobility and providing early detection of symptomatic patients. In this paper, a smart real-time image processing framework converged with a non-contact thermal temperature screening module was implemented. The proposed framework comprised of three modules , smart temperature screening system, tracking infection footprint, and monitoring social distancing policies. This was accomplished by employing Histogram of Oriented Gradients (HOG) transformation to identify infection hotspots. Further, Haar Cascade and local binary pattern histogram (LBPH) algorithms were used for real-time facial recognition and enforcing social distancing policies. In order to manage the redundant video frames generated at the local computing device, two holistic models, namely, event-triggered video framing (ETVF) and real-time video framing (RTVF) have been deduced, and their respective processing costs were studied for different arrival rates of the video frame. It was observed that the proposed ETVF approach outperforms the performance of RTVF by providing optimal processing costs resulting from the elimination of redundant data frames. Results corresponding to autocorrelation analysis have been presented for the case study of India pertaining to the number of confirmed COVID-19 cases, recovered cases, and deaths to obtain an understanding of epidemiological spread of the virus. Further, choropleth analysis was performed for indicating the magnitude of COVID-19 spread corresponding to different regions in India.
近年来，已采取了多种策略来尽量减少2019新型冠状病毒病（COVID-19）的传播。最近的一些技术突破，如基于无人机的跟踪系统，专注于设计特定的动态方法来管理公共交通出行，并对有症状的患者进行早期检测。在本文中，实现了一个与非接触式热温度筛查模块相结合的智能实时图像处理框架。所提出的框架由三个模块组成，即智能温度筛查系统、追踪感染足迹和监测社交距离政策。这是通过使用方向梯度直方图（HOG）变换来识别感染热点来实现的。此外，Haar级联和局部二值模式直方图（LBPH）算法用于实时面部识别和执行社交距离政策。为了管理本地计算设备生成的冗余视频帧，推导了两种整体模型，即事件触发视频帧（ETVF）和实时视频帧（RTVF），并研究了它们在不同视频帧到达率下各自的处理成本。结果表明，所提出的ETVF方法通过消除冗余数据帧提供了最佳处理成本，从而优于RTVF的性能。针对印度的案例研究，给出了与确诊COVID-19病例数、康复病例数和死亡数相关的自相关分析结果，以了解该病毒的流行病学传播情况。此外，还进行了分级统计图分析，以表明印度不同地区COVID-19传播的程度。
