Automated Face Mask Detection using Artificial Intelligence and Video Surveillance Management

Surveillance camera has become an essential, ubiquitous technology in people's daily lives, whether applicable for home surveillance or extended to public workplace detection. The importance of the camera is irreplaceable in terms of the agent for an enclosed system to function correctly. The goal of ubiquitous computing is to keep different devices or technology communicating seamlessly, allowing them to expand to other areas instead of limiting it to one device. However, many research papers have been released on how the camera can aid in the current situation where COVID-19 is still raging worldwide, especially in crowded places. This paper aims to suggest a method by which surveillance cameras on the university campus can automatically detect student face mask status and notify them. Alongside that, this concept of applying a video management system within the university campus will assist in the automation of invigilating the student's daily mask status from the number of embedded surveillance cameras around the campus. © 2023 IEEE.

Object detection and activity recognition in video surveillance using neural networks

Purpose: This paper aims to implement and extend the You Only Live Once (YOLO) algorithm for detection of objects and activities. The advantage of YOLO is that it only runs a neural network once to detect the objects in an image, which is why it is powerful and fast. Cameras are found at many different crossroads and locations, but video processing of the feed through an object detection algorithm allows determining and tracking what is captured. Video Surveillance has many applications such as Car Tracking and tracking of people related to crime prevention. This paper provides exhaustive comparison between the existing methods and proposed method. Proposed method is found to have highest object detection accuracy. Design/methodology/approach: The goal of this research is to develop a deep learning framework to automate the task of analyzing video footage through object detection in images. This framework processes video feed or image frames from CCTV, webcam or a DroidCam, which allows the camera in a mobile phone to be used as a webcam for a laptop. The object detection algorithm, with its model trained on a large data set of images, is able to load in each image given as an input, process the image and determine the categories of the matching objects that it finds. As a proof of concept, this research demonstrates the algorithm on images of several different objects. This research implements and extends the YOLO algorithm for detection of objects and activities. The advantage of YOLO is that it only runs a neural network once to detect the objects in an image, which is why it is powerful and fast. Cameras are found at many different crossroads and locations, but video processing of the feed through an object detection algorithm allows determining and tracking what is captured. For video surveillance of traffic cameras, this has many applications, such as car tracking and person tracking for crime prevention. In this research, the implemented algorithm with the proposed methodology is compared against several different prior existing methods in literature. The proposed method was found to have the highest object detection accuracy for object detection and activity recognition, better than other existing methods. Findings: The results indicate that the proposed deep learning–based model can be implemented in real-time for object detection and activity recognition. The added features of car crash detection, fall detection and social distancing detection can be used to implement a real-time video surveillance system that can help save lives and protect people. Such a real-time video surveillance system could be installed at street and traffic cameras and in CCTV systems. When this system would detect a car crash or a fatal human or pedestrian fall with injury, it can be programmed to send automatic messages to the nearest local police, emergency and fire stations. When this system would detect a social distancing violation, it can be programmed to inform the local authorities or sound an alarm with a warning message to alert the public to maintain their distance and avoid spreading their aerosol particles that may cause the spread of viruses, including the COVID-19 virus. Originality/value: This paper proposes an improved and augmented version of the YOLOv3 model that has been extended to perform activity recognition, such as car crash detection, human fall detection and social distancing detection. The proposed model is based on a deep learning convolutional neural network model used to detect objects in images. The model is trained using the widely used and publicly available Common Objects in Context data set. The proposed model, being an extension of YOLO, can be implemented for real-time object and activity recognition. The proposed model had higher accuracies for both large-scale and all-scale object detection. This proposed model also exceeded all the other previous methods that were compared in extending and augmenting the object detection to activity recognition. The proposed model resulted in the highest accuracy for car crash detection, fall detection and social distancing detection. © 2023, Emerald Publishing Limited.

Development of a Video Surveillance web Application for the use of Biosafety Equipment with Image Recognition for Workspaces

With the arrival of Covid-19, several preventive measures were implemented to limit the spread of this virus. Among these measures is the use of masks, both in open and closed public spaces. This measure has forced commercial establishments, workplaces, schools, hospitals, to maintain constant vigilance, upon entering their facilities, of the proper use of the mask, which should completely cover the nose, mouth and chin. However, this manual control is tedious and ineffective since most of the population is not able to correctly identify when a person has the mask on properly, with high error rates in the manual detection of the correct use of the mask according to surveys carried out. For this reason, this work proposes the automation of the detection of the proper use of the mask at the entrance to the work areas, also providing a follow-up panel of the recorded incidents. The effectiveness of the proposal was evaluated through the detection and categorization of a data set of more than 3000 images, resulting in an accuracy of 98.6%. © 2022 IEEE.

Realtime Risk Monitoring System using Convolutional Neural Network for COVID19 Pandemic

The COVID19 pandemic has significantly changed the lifestyle of billions of people across the globe. It has greatly affected almost all sectors of business, industry and public life. As per the WHO's guidelines, wearing a face mask has become the new compulsory and precautionary measures for everyone. Currently, all the public and private service providers will expect their stakeholders to wear face mask in an appropriate way to avail any services. Therefore, detection of face mask at public places is a crucial task to help the society to overcome current pandemic. This paper presents a unique approach to not only detect face mask but also calculate the risk of getting infected by COVID-19 using machine learning algorithms. The proposed model detects the various faces present in an input video, identifies if it has a mask present or not. If the mask is not detected, the model calculates the risk of human being getting infected based on their age. Finally, the model generates the output and provides analysis based on the real time data it has processed. As a real-time surveillance system, the model can also classify a face when a person is moving in the live video. The proposed method attained a highest accuracy of 99.57 % against standard datasets under study. The authors experimented and explored various Convolutional Neural Network models like DenseNet, MobileNet_V2, Inception_V3 and YOLO_V4 find the best model, detecting the presence of masks accurately without causing over-fitting. © 2022 IEEE.

Automated Physical Distance Estimation and Crowd Monitoring Through Surveillance Video.

The contagious Corona Virus (COVID-19) transmission can be reduced by following and maintaining physical distancing (also known as COVID-19 social distance). The World Health Organisation (WHO) recommends preventing COVID-19 from spreading in public areas. On the other hand, people may not be maintaining the required 2-m physical distance as a mandated safety precaution in shopping malls and public places. The spread of the fatal disease may be slowed by an active monitoring system suitable for identifying distances between people and alerting them. This paper introduced a deep learning-based system for automatically detecting physical distance using video from security cameras. The proposed system introduced the TH-YOLOv5 for object detection and classification and Deepsort for tracking the detected people using bounding boxes from the video. TH-YOLOv5 included another prediction head to identify objects of varying sizes. The original prediction heads are then replaced with Transformer Heads (TH) to investigate the prediction capability of the self-attention mechanism. Then, we include the convolutional block attention model (CBAM) to identify attention areas in settings with dense objects. Pairwise L2 vectorized normalization was utilized to generate a three-dimensional feature space for tracking physical distances and the violation index, determining the number of individuals who follow the distance rules. We use the MS COCO and HumanCrowd, CityPersons, and Oxford Town Centre (OTC) data sets for training and testing. Experimental results demonstrate that the proposed system obtained a weighted mAP score of 89.5% and an FPS score of 29; both are computationally comparable.

K-Zones: a Machine Learning-Based System to Estimate Social Distancing Violations During Pandemic Eras

The outbreak of pandemics adversely influences various aspects of people's lives, including economies, education, careers, and social relations. Therefore, many authorities worldwide resort to imposing social distancing regulations to flatten the curve of new confirmed cases. This paper proposes a Machine Learning-based social distancing violation detection system. Unlike many contributions in the literature that use pairwise distance computation running in quadratic execution time, this paper introduces a novel technique that runs in linear time. The solution is considered a Video Surveillance System, and the experimental results show how the system effectively detects not only social distancing violations but also the severity of those violations. © 2022 IEEE.

MONITORED ISOLATION ROOM: AN ALTERNATIVE SOLUTION DURING COVID-19 CRISIS

Background: Acute stroke unit care is proven to reduce mortality and morbidity. During the COVID-19 crisis, we must guarantee the provision of acute stroke care and optimize care protocols to reduce the risk of SARS-CoV-2 infection and rationalize the use of hospital resources. Our hospital developed an adapted protocol which includes individual isolation room equipped with a monitor connected to the central monitoring unit of the stroke unit, and a camera that allowed patient supervision from the control nursing unit for stroke patients with suspected or confirmed COVID-19 infection. We present a descriptive study of our experience. Methods: Observational, extensive, and transversal study. Patients admitted to the monitored isolation room of our stroke unit between November-2020 to december-2021. Results: 201 patients, 51,7.% women. 76.1% ischemic stroke, of which 10% had been treated with thrombolysis and 2.9% with thrombectomy. 90.3% without infectious symptoms. In 6.2% the Covid infection was known before their arrival at the emergency room, in 3 patient it was detected in the emergency room and in 4 during their stay in the isolation room. No contagions were detected within the stroke unit after the introduction of this measure. Only 10,1% of the patients stayed in the room for more than 24 hours. Camera detect care needs in 22% of the patients. The destination at discharge was conventional stroke unit in 83.4% of the patients. Conclusions: an isolation room monitored and controlled by video surveillance is an effective alternative to prevent infections in the stroke unit.

Integrating Deep Learning-Based IoT and Fog Computing with Software-Defined Networking for Detecting Weapons in Video Surveillance Systems.

Due to the widespread proliferation of multimedia traffic resulting from Internet of Things (IoT) applications and the increased use of remote multimedia-based applications, as a consequence of COVID-19, there is an urgent need to develop intelligent adaptive techniques that improve the Quality of Service (QoS) perceived by end-users. In this work, we investigate the integration of deep learning techniques with Software-Defined Network (SDN) architecture to support delay-sensitive applications in IoT environments. Weapon detection in real-time video surveillance applications is deployed as our case study upon which multiple deep learning-based models are trained and evaluated for detection using precision, recall, and mean absolute precision. The deep learning model with the highest performance is then deployed within a proposed artificial intelligence model at the edge to extract the first detected video frames containing weapons for quick transmission to authorities, thus helping in the early detection and prevention of different kinds of crimes, and at the same time decreasing the bandwidth requirements by offloading the communication network from massive traffic transmission. Performance improvement is achieved in terms of delay, throughput, and bandwidth requirements by dynamically programming the network to provide different QoS based on the type of offered traffic and current traffic load, and based on the destination of the traffic. Performance evaluation of the proposed model was carried out using the mininet emulator, which revealed improvement of up to 75.0% in terms of average throughput, up to 14.7% in terms of mean jitter, and up to 32.5% in terms of packet loss.

A Deep-Learning Based Automated COVID-19 Physical Distance Measurement System Using Surveillance Video

The contagious Corona Virus (COVID-19) transmission can be reduced by following and maintaining physical distancing (also known as COVID-19 social distance). The World Health Organisation (WHO) recommends it to prevent COVID-19 from spreading in public areas. On the other hand, people may not be maintaining the required 2-m physical distance as a mandated safety precaution in shopping malls and public places. The spread of the fatal disease may be slowed by an active monitoring system suitable for identifying distances between people and alerting them. This paper introduced a deep learning-based system for automatically detecting physical distance using video from security cameras. The proposed system employed the fine-tuning YOLO v4 for object detection and classification and Deepsort for tracking the detected people using bounding boxes from the video. Pairwise L2 vectorized normalization was utilized to generate a three-dimensional feature space for tracking physical distances and the violation index, determining the number of individuals who follow the distance rules. For training and testing, we use the MS COCO and Oxford Town Centre (OTC) datasets. We compared the proposed system to two well-known object detection models, YOLO v3 and Faster RCNN. Our method obtained a weighted mAP score of 87.8% and an FPS score of 28;both are computationally comparable. © 2022, Springer Nature Switzerland AG.

Surveillance in Schools across Europe: A New Phenomenon in Light of the COVID-19 Pandemic? The Cases of Greece and France

Surveillance technology is more and more used in educational environments, which results in mass privacy violations of kids and, thus, the processing of huge amount of children’s data in the name of safety. Methodology used is doctrinal, since the focus of this research was given in the implementation of the legal doctrine of data protection law in the educational environments. More than that, the cases of Greece and France regarding the use of surveillance technologies in schools are carefully studied in this article. Privacy risks that both children and educators are exposed to are underlined. In these terms, this research paper focuses on the proper implementation of the European data protection framework and the role of Data Protection Authorities as control mechanisms, so that human rights risks from the perspective of privacy and data protection to be revealed, and the purposes of the use of such technologies to be evaluated. This study is limited in the legal examination of the European General Data Protection Regulation, and its implementation in the legal orders of Greece and France, and practice pertaining to the case studies of Greece and France respectively. © 2022 The Author(s).

Social Distancing Surveillance System via Inverse Perspective Mapping and Fixed-point Quantization

During the Coronavirus Disease 2019 (COVID-19) pandemic, many countries have introduced the social distancing policy in public areas to stop the spread of disease by maintaining a physical distance between people. This paper proposes an Artificial Intelligence (AI)-powered social distancing surveillance system that can detect pedestrians through video surveillance and monitor the social distance between them via Inverse Perspective Mapping (IPM) in real-time. The proposed system was deployed on the devices located at the network edge such as IoT devices and mobile devices to enable real-time response with low data transmission latency. To bypass the restriction on the computational and memory capacity for the edge devices, the proposed system was optimized through fixed-point quantization. From the evaluation results, the optimized models are almost 4 times smaller as compared to the original models. The best trade-off between speed and accuracy can be achieved with a 27.1% improvement in speed and 2% degradation in accuracy.

COVID-19 Social Distancing Surveillance System

The pandemic from COVID-19 impinged our day-to-day lives and wreaked havoc upon many sectors in our society. This worldwide pandemic, which had its onset in January 2020, has forced us to reconsider our perception of what “normal” should be. While there’s no official cure yet, various vaccines have been rolled out and are expected to take effect soon. However, the efficacy of vaccines has been a debatable issue. Thus, the most effective way to battle this situation would be to strictly follow the precautionary measures advised by the governing authorities. Wearing mask and following the social distancing norms are considered as one of the most effective ways to control the spread of infection [1]. However, this new normal becomes difficult to implement as many people tend not to follow social distancing. While it is difficult to check whether people are following social distancing, we propose a solution which would come in handy in such circumstances and would hasten the process of contact tracing in comparison to manual inspection. In this study, we strive to present a video surveillance model, which would allow the detection of social distancing between people based on object detection and tracking algorithms. The specific algorithm used in our study for object detection is the YOLO algorithm and monitoring the distance between any two persons is done using a technique called Perspective Transformation. The proposed method shows promising results which could be implemented as a surveillance system for monitoring social distancing. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Facemask Detection to Prevent COVID-19 Using YOLOv4 Deep Learning Model

The on-going global Covid-19 pandemic has impacted everyone's life. World Health organization (WHO) and Governments all over world have found that social distancing and donning a mask in public places has been instrumental in reducing the rate of COVID-19 transmission. Stepping out of homes in a face mask is a social obligation and a law mandate that is often violated by people and hence a face mask detection model that is accessible and efficient will aid in curbing the spread of disease. Detecting and identifying a face mask on an individual in real time can be a daunting and challenging task but using deep learning and computer vision, establish tech-based solutions that can help combat COVID-19 pandemic. In this paper, YOLOv4 deep learning model is designed and applied deep transfer learning approach to create a face mask detector which can be used in real time. GPU used was Google Collab to run the simulations and to draw inferences. Proposed implementation considered three types of data as input such as image dataset, video dataset and real time data for face mask detection. Performance parameters are tabulated and obtained mean average precision of 0.86, F1 score 0.77 for image dataset, 90 % accuracy for video dataset. And real time face mask detector with accuracy of 95%, it is successfully able to identify a person with and without facemask and report if they are wearing a face mask or not. © 2022 IEEE.

Social Distancing Violation Detection Using Pre-Trained Object Detection Models

The outbreak of COVID-19 pandemic has resulted in a devastating impact all around the world. The social distancing protocol has become the compulsory preventive measures in many countries for the purpose of avoiding physical contacts with other persons. This paper presents a drone based surveillance approach that uses Computer Vision and Deep Learning based techniques to check if two persons are violating social distancing norms. The implementation has been done using a Raspberry Pi based Drone where the camera captures the video from a height and detects the persons who are in near proximity with the help of an object detection algorithm and computes the distance between two persons in order to check whether they lie near to each other. The Euclidean distance between the two persons is calculated which is then compared with the given margin distance. If the distance is found to be below the margin., local authorities or law enforcement agencies can be notified through an automatically generated email alert. © 2021 IEEE.

Object Detection for COVID Rules Response and Crowd Analysis

Understanding the hotspots attracting massive crowds is a huge necessity during this pandemic times. The knowledge of analyzing crowds will help to plan and avoid the spread of the virus to a large extent by identifying exact hotspots. Understanding where the crowds flock and whether they are following the guidelines or not will help in taking appropriate actions, allotting concerned personnel in advance, and closing of areas which are at higher risks can be advantageous. In order to realize the situation, real-time analysis of the pandemic rules like social distancing, wearing masks is necessary. This paper proposes the use of video surveillance and provides a combined application to check the factors necessary during crowd situations as per rules set by the Government. This work uses python as a coding language, and YOLOv4 algorithm along with various libraries like darknet to improve video and image analysis for the identification of exact requirements. This work also uses Cuda software and Cudnn library for the acceleration of processing. The paper proposes importantly, counting people passing through a particular area, detecting whether people are following social distancing, detecting if the participants are wearing a mask, and counting the number of vehicles passing through an area. The knowledge of analyzing crowds will help to plan and avoid the spread of the virus to a large extent by identifying exact hotspots. All the applications are connected to the graphical user interface (GUI) and depending on the input each application proposed considers different analysis. The types of input are image, video, image directory and live feed are considered to obtain better results. © 2021 IEEE.

A novel algorithm for mask detection and recognizing actions of human.

Face recognition has become a significant challenge today since an increasing number of individuals wear masks to avoid infection with the novel coronavirus or Covid-19. Due to its rapid proliferation, it has garnered growing attention. The technique proposed in this chapter seeks to produce unconstrained generic actions in the video. Conventional anomaly detection is difficult because computationally expensive characteristics cannot be employed directly, owing to the necessity for real-time processing. Even before activities are completely seen, they must be located and classified. This paper proposes an expanded Mask R-CNN (Ex-Mask R-CNN) architecture that overcomes these issues. High accuracy is achieved by using robust convolutional neural network (CNN)-based features. The technique consists of two steps. First, a video surveillance algorithm is employed to determine whether or not a human is wearing a mask. Second, Multi-CNN forecasts the frame's suspicious conventional abnormality of people. Experiments on tough datasets indicate that our approach outperforms state-of-the-art online traditional detection of anomaly systems while maintaining the real-time efficiency of existing classifiers.

Sparse Crowd Flow Analysis of Tawaaf of Kaaba During the COVID-19 Pandemic

The advent of the COVID-19 pandemic has adversely affected the entire world and has put forth high demand for techniques that remotely manage crowd-related tasks. Video surveillance and crowd management using video analysis techniques have significantly impacted today's research, and numerous applications have been developed in this domain. This research proposed an anomaly detection technique applied to Umrah videos in Kaaba during the COVID-19 pandemic through sparse crowd analysis. Managing the Kaaba rituals is crucial since the crowd gathers from around the world and requires proper analysis during these days of the pandemic. The Umrah videos are analyzed, and a system is devised that can track and monitor the crowd flow in Kaaba. The crowd in these videos is sparse due to the pandemic, and we have developed a technique to track the maximum crowd flow and detect any object (person) moving in the direction unlikely of the major flow. We have detected abnormal movement by creating the histograms for the vertical and horizontal flows and applying thresholds to identify the non-majority flow. Our algorithm aims to analyze the crowd through video surveillance and timely detect any abnormal activity to maintain a smooth crowd flow in Kaaba during the pandemic. © 2022 Tech Science Press. All rights reserved.

Quick Response in the SARS-COV-2 Context. How to Build up a Patient Video Monitoring System in Two days

The spike in the number of cases of SARS-COV-2 in Cluj county in March 2020 brought up a set of challenges for the affected medical personnel due to this atypical context, tied to the rapid evolution of pandemic, rather old infrastructure requiring upgrade, limited expertise with a similar situation. To avoid unnecessary exposure of physicians and nurses to the SARS-COV-2 infected patients, a video surveillance system was deployed. This was not a trivial task because of the various technical issues related to infrastructure, short development time and other specific constraints. This paper presents how such solution got implemented, using a mixture of off-the-shelf equipment, combined with flexible graphical programming environment used in industry and academia. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

A Social Distance Estimation and Crowd Monitoring System for Surveillance Cameras.

Social distancing is crucial to restrain the spread of diseases such as COVID-19, but complete adherence to safety guidelines is not guaranteed. Monitoring social distancing through mass surveillance is paramount to develop appropriate mitigation plans and exit strategies. Nevertheless, it is a labor-intensive task that is prone to human error and tainted with plausible breaches of privacy. This paper presents a privacy-preserving adaptive social distance estimation and crowd monitoring solution for camera surveillance systems. We develop a novel person localization strategy through pose estimation, build a privacy-preserving adaptive smoothing and tracking model to mitigate occlusions and noisy/missing measurements, compute inter-personal distances in the real-world coordinates, detect social distance infractions, and identify overcrowded regions in a scene. Performance evaluation is carried out by testing the system's ability in person detection, localization, density estimation, anomaly recognition, and high-risk areas identification. We compare the proposed system to the latest techniques and examine the performance gain delivered by the localization and smoothing/tracking algorithms. Experimental results indicate a considerable improvement, across different metrics, when utilizing the developed system. In addition, they show its potential and functionality for applications other than social distancing.

Management of safe distancing on construction sites during COVID-19: A smart real-time monitoring system.

The outbreak of Coronavirus Disease 2019 (COVID-19) poses a great threat to the world. One mandatory and efficient measure to prevent the spread of COVID-19 on construction sites is to ensure safe distancing during workers' daily activities. However, manual monitoring of safe distancing during construction activities can be toilsome and inconsistent. This study proposes a computer vision-based smart monitoring system to automatically detect worker breaching safe distancing rules. Our proposed system consists of three main modules: (1) worker detection module using CenterNet; (2) proximity determination module using Homography; and (3) warning alert and data collection module. To evaluate the system, it was implemented in a construction site as a case study. This study has two key contributions: (1) it is demonstrated that monitoring of safe distancing can be automated using our approach; and (2) CenterNet, an anchorless detection model, outperforms current state-of-the-art approaches in the real-time detection of workers.