A Social Distance Monitoring Method Based on Improved YOLOv4 for Surveillance Videos

Social distance monitoring is of great significance for public health in the era of COVID-19 pandemic. However, existing monitoring methods cannot effectively detect social distance in terms of efficiency, accuracy, and robustness. In this paper, we proposed a social distance monitoring method based on an improved YOLOv4 algorithm. Specifically, our method constructs and pre-processes a dataset. Afterwards, our method screens the valid samples and improves the K-means clustering algorithm based on the IoU distance. Then, our method detects the target pedestrians using a trained improved YOLOv4 algorithm and gets the pedestrian target detection frame location information. Finally, our method defines the observation depth parameters, generates the 3D feature space, and clusters the offending aggregation groups based on the L2 parametric distance to finally realize the pedestrian social distance monitoring of 2D video. Experiments show that the proposed social distance monitoring method based on improved YOLOv4 can accurately detect pedestrian target locations in video images, where the pre-processing operation and improved K-means algorithm can improve the pedestrian target detection accuracy. Our method can cluster the offending groups without going through calibration mapping transformation to realize the pedestrian social distance monitoring of 2D videos. © 2023 World Scientific Publishing Company.

An Efficient Medicine Identification and Delivery System Based on Mobile Manipulation Robot

In recent years, with the emergence of COVID-19, the shortage of medical resources has become increasingly obvious. However, current environments such as hospital wards still require a large number of medical staff to deliver medicines. In this paper, we propose a mobile robot that can complete medicine grabbing and delivery in a hospital ward scenario. First, a lightweight neural network is built to improve the detection efficiency of Faster R-CNN algorithm for boxed medicine. Then, the pose of the robotic arm grasping the pill box is determined by point cloud matching to control the mechanical grasping of the pill box. Finally, a discomfort function representing the collision risk between the robot and the pedestrian is incorporated into the Risk-RRT algorithm to improve the navigation performance of the algorithm. By building a real experimental platform, the experiments verify the performance of our proposed medicine delivery robot system. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2022.

Detection of Face Wearing Mask Based on AdaBoost and YCrCb

At the beginning of 2020, COVID-19 broke out and swept the world. Wearing masks remains an important means of preventing epidemics. Many scholars have developed and studied mask wearing detection based on YOLO algorithm, and have made some achievements. AdaBoost algorithm has the advantages of high precision and low complexity, and is also suitable for solving this problem. This paper uses OpenCV to propose a face detection algorithm based on AdaBoost. This algorithm is based on face detection, including initialization of background estimation example, background subtraction preprocessing, obtaining eye position, face detection and other steps. LBP features are used as the training basis of the classifier. The trained classifier is generated and used as a function in the mask detection algorithm. At present, there are two problems in the research of mask wearing detection: first, only consider whether the tested object wears a mask, but not analyze the non-standard wearing of masks;Secondly, due to the influence of light and other external environments, the real-time detection effect of targets in complex scenes changes greatly. In view of the above problems, this paper adopts the following methods to solve them: pre-processing the image to reduce noise, light spots and other external environmental interference;For the case that the mask is not standardized, the condition that the mask covers the nose and mouth shall be detected. Finally, the Adaboost algorithm for facial mask wearing detection is obtained. Experiments show that the algorithm has high adaptability, robustness and accuracy, and can be used to promote the development of epidemic prevention. © 2022 IEEE.

Dormitory Food Delivery Smart Car

This paper designs a smart car that can automatically deliver meals in dormitories, isolated hotels and other scenarios. This system uses i.MX RT1064 as the main controller, and completes the route tracking and room number recognition of the smart car through the MT9V034 camera and the OpenART mini visual sensor module respectively. The target detection method is the SSD algorithm in the one-stage method. After optimization, the recognition rate is as high as 90%, which can successfully complete the meal delivery task. This system greatly reduces the risk of human-to-human contact, reduces the probability of contracting COVID-19, and contributes to epidemic prevention and control measures to minimize risks. © 2022 SPIE. All rights reserved.

Automatic Detection of Dental Lesions Based on Deep Learning

Objective Since the outbreak of COVID-19, many hospitals have become overloaded with patients seeking examination, resulting in an imbalance between medical staff and patients. These high concentrations of people in hospital settings not only aggravate the risk of cross-infection among patients, but also stall the public medical system. Consequently, mild and chronic conditions cannot be treated effectively, and eventually develop into serious diseases. Therefore, the use of deep learning to accurately and efficiently analyze X-ray images for diagnostic purposes is crucial in alleviating the pressure on medical institutions during epidemics. The method developed in this study accurately detects dental X-ray lesions, thus enabling patients to self-diagnose dental conditions. Methods The method proposed in this study employs the YOLOV5 algorithm to detect lesion areas on digital X-ray images and optimize the network model's parameters. When hospitals and medical professionals collect and label training data, they use image normalization to enhance the images. Consequently, in combination with the network environment, parameters were adjusted into four modules in the YOLOV5 algorithm. In the Input module, Mosaic data enhancement and adaptive anchor box algorithms are used to generate the initial box. The focus component was added to the Backbone module, and a CSP structure was implemented to determine the image features. When the obtained image features are input into the Backbone module, the FPN and PAN structures are used to realize feature fusion. Subsequently, GIOU_Loss function is applied to the Head moudule, and NMS non-maximum suppression is used to generate a regression of results. Results and Discussions The proposed YOLOV5-based neural network yields satisfactory training and testing results. The training algorithm produced a recall rate of 95%, accuracy rate of 95%, and F1 score of 96%. All evaluation criteria are higher than those of the target detection algorithms of SSD and Faster-RCNN (Table 1). The network converges to smoothness after loss is reduced in the training process (Fig. 6), which proves that the network successfully learns the necessary features. Thus, the difference between predicted and real values is very small, which indicates good model performance. The mAP value of network training is 0.985 (Fig. 7), which proves that the network training meets the research requirements. Finally, an observation of the visualized thermodynamic diagram reveals that the network's region of interest matches the target detection region (Fig. 8). Conclusions This study proposes the use of the YOLOV5 algorithm for detecting lesions in dental X-ray images, training and testing on the dataset, modifying the network's nominal batch size, selecting an appropriate optimizer, adjusting the weight parameters, and modifying the learning rate attenuation strategy. The model's training results were compared with those of algorithms used in previous studies. Finally, the effect of feature extraction was analyzed after the thermodynamic diagram was visualized. The experimental results show that the algorithm model detects lesion areas with an accuracy rate of more than 95%, making it an effective autonomous diagnostic tool for patients. © 2022 Science Press. All rights reserved.

Lightweight Real-Time Mask Wearing Detection Algorithm Based on YOLOv3

The novel coronavirus has broken out from 2019 and quickly become a global pandemic. It spreads rapidly from person to person through droplets, aerosols and other carriers. In order to prevent the spread of the virus, people must wear masks when entering and leaving public places and taking public transport to reduce the risk of virus transmission. How to detect the wearing of masks in public places and other natural environments has become a new research problem. This paper proposes a lightweight deep neural network (E-YOLO) to realize mask wearing detection in real-time scenarios. E-YOLO improves the general target detection YOLOv3 algorithm by the follows methods. Firstly, the Efficient-Net series B2 backbone feature extraction network is used to replace the original Darknet53 feature extraction network. Combined with the spatial pyramid pooling module and the bidirectional feature pyramid structure to enrich the semantic information of the feature layer, so as to achieve the balance between the speed and accuracy of target detection in real-time scenes. The experimental results show that, compared with the YOLOv3 algorithm, the E-YOLO algorithm possesses the same accuracy and speed. The network floating point calculation amount is one twelfth of YOLOv3, and the model size is only one quarter of YOLOv3, which is more suitable for resource-constrained platforms. © 2021 IEEE.