Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9824328
Hong Zhou, Guoqiang Chen
Community structure is an important property in complex networks. Community detection can be formulated as optimization problem. The single objective commonly used in the analysis of static networks is often powerless in the face of conflicting optimization demands. In this paper, a multi-objective whale optimization based community detection algorithm (MOWOCD) is proposed, MOWOCD can optimize KKM and RC simultaneously. Experiments on real life networks show that MOWOCD can get effective results.
{"title":"A multiobjective optimization whale optimization based community detection algorithm","authors":"Hong Zhou, Guoqiang Chen","doi":"10.1109/cvidliccea56201.2022.9824328","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9824328","url":null,"abstract":"Community structure is an important property in complex networks. Community detection can be formulated as optimization problem. The single objective commonly used in the analysis of static networks is often powerless in the face of conflicting optimization demands. In this paper, a multi-objective whale optimization based community detection algorithm (MOWOCD) is proposed, MOWOCD can optimize KKM and RC simultaneously. Experiments on real life networks show that MOWOCD can get effective results.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"97 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86492492","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9824180
Zihao Yu
Scene classification is a popular and important question in computer vision and has been developed in different areas. Applying computer vision to artworks has become a popular topic in recent years. However, the traditional random sampling to identify the artworks through machine learning requires a large data set and, therefore, a higher cost to get a solid result. This paper compares random sampling and active learning (uncertainty sampling) performance using a data set (8446 paintings) of the 50 most influential painters in Europe from the 13th to the 20th century. and then propose that the active learning strategy can build a stronger model that requires smaller data sets. The active learning model can be further improved through training in larger data sets and applied in the artwork recognition for artificial intelligence..
{"title":"Discerning Art Works through Active Machine Learning","authors":"Zihao Yu","doi":"10.1109/cvidliccea56201.2022.9824180","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9824180","url":null,"abstract":"Scene classification is a popular and important question in computer vision and has been developed in different areas. Applying computer vision to artworks has become a popular topic in recent years. However, the traditional random sampling to identify the artworks through machine learning requires a large data set and, therefore, a higher cost to get a solid result. This paper compares random sampling and active learning (uncertainty sampling) performance using a data set (8446 paintings) of the 50 most influential painters in Europe from the 13th to the 20th century. and then propose that the active learning strategy can build a stronger model that requires smaller data sets. The active learning model can be further improved through training in larger data sets and applied in the artwork recognition for artificial intelligence..","PeriodicalId":23649,"journal":{"name":"Vision","volume":"66 1","pages":"1002-1006"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82904925","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9824618
Qian Wang, Yaoran Huo
A secure single-pixel imaging (SPI) method which based on 2D Logistic map is proposed in this paper. A 2D Logistic map combined with its key and the asymmetric ratio is utilized to generate the asymmetric measurement matrices and encrypt the measurements. Compared with the existing secure SPI methods, the proposed method achieves the same security and the asymmetric measurement matrices suppress the noise effectively with a better imaging quality. The key of the chaos system combined with the asymmetric ratio in the asymmetric measurement matrices are transmitted in the secret channel instead of the measurement matrices. Based on the proposed strategy, transmitted data size is reduced. Simulation results show that our method achieves advantages in image quality and security. The proposed strategy can also drop the transmitted data size and computational complexity.
{"title":"Single-pixel imaging encryption based on 2D coupled Logistic mapping","authors":"Qian Wang, Yaoran Huo","doi":"10.1109/cvidliccea56201.2022.9824618","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9824618","url":null,"abstract":"A secure single-pixel imaging (SPI) method which based on 2D Logistic map is proposed in this paper. A 2D Logistic map combined with its key and the asymmetric ratio is utilized to generate the asymmetric measurement matrices and encrypt the measurements. Compared with the existing secure SPI methods, the proposed method achieves the same security and the asymmetric measurement matrices suppress the noise effectively with a better imaging quality. The key of the chaos system combined with the asymmetric ratio in the asymmetric measurement matrices are transmitted in the secret channel instead of the measurement matrices. Based on the proposed strategy, transmitted data size is reduced. Simulation results show that our method achieves advantages in image quality and security. The proposed strategy can also drop the transmitted data size and computational complexity.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"3 1","pages":"683-686"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90148422","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In agricultural productivity, detecting plant pests and diseases is extremely crucial. This research studies images of plant leaf pests and diseases from an unsupervised perspective to solve the problem that existing plant leaf disease datasets are difficult to acquire and include few types of diseases, and they cannot find the defective parts of leaves. This paper utilizes the idea of image restoration and uses a deep learning correlation model to detect and localize the abnormal regions of plant leaves. The experimental results show that the img_AUCROC and pixel_AUCROC level anomaly detection and localization achieve good results, which bring influence and reference to other peers.
{"title":"Application research of plant leaf pests and diseases base on unsupervised learning","authors":"Mingjing Pei, Min Kong, MaoSheng Fu, Xiancun Zhou, Zusong Li, Jieru Xu","doi":"10.1109/cvidliccea56201.2022.9824321","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9824321","url":null,"abstract":"In agricultural productivity, detecting plant pests and diseases is extremely crucial. This research studies images of plant leaf pests and diseases from an unsupervised perspective to solve the problem that existing plant leaf disease datasets are difficult to acquire and include few types of diseases, and they cannot find the defective parts of leaves. This paper utilizes the idea of image restoration and uses a deep learning correlation model to detect and localize the abnormal regions of plant leaves. The experimental results show that the img_AUCROC and pixel_AUCROC level anomaly detection and localization achieve good results, which bring influence and reference to other peers.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"2 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"83038367","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9825055
Wei-wu Guo, Nanbo Shen, Tingjuan Zhang
Pedestrian detection in a crowded environment is challenging for vehicle intelligent driving systems. At present, pedestrian detection algorithms have achieved great performance in detecting well-separated figures. However, pedestrians are generally overlapped in crowded scenes, resulting in slow detection speed, low detection accuracy, and poor robustness in pedestrian detection technology. In this paper, the YOLOv5 algorithm is used for pedestrian detection. In the aspect of data pretreatment, Mosaic data enhancement, unified image size, adaptive anchor frame calculation, and other processing are carried out for data.YOLOv5 can detect targets at multiple scales, and CIOU_Loss and DIOU_nms are applied to the YOLOv5 algorithm. It can improve the recognition ability of the occlusion target and has a good detection effect on the detection of the occlusion pedestrian target through the training network of amplified data set. The verification experiment shows that the pedestrian detection model based on YOLOv5 has great detection accuracy and recall rate in detecting covered pedestrians.
{"title":"Overlapped Pedestrian Detection Based on YOLOv5 in Crowded Scenes","authors":"Wei-wu Guo, Nanbo Shen, Tingjuan Zhang","doi":"10.1109/cvidliccea56201.2022.9825055","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9825055","url":null,"abstract":"Pedestrian detection in a crowded environment is challenging for vehicle intelligent driving systems. At present, pedestrian detection algorithms have achieved great performance in detecting well-separated figures. However, pedestrians are generally overlapped in crowded scenes, resulting in slow detection speed, low detection accuracy, and poor robustness in pedestrian detection technology. In this paper, the YOLOv5 algorithm is used for pedestrian detection. In the aspect of data pretreatment, Mosaic data enhancement, unified image size, adaptive anchor frame calculation, and other processing are carried out for data.YOLOv5 can detect targets at multiple scales, and CIOU_Loss and DIOU_nms are applied to the YOLOv5 algorithm. It can improve the recognition ability of the occlusion target and has a good detection effect on the detection of the occlusion pedestrian target through the training network of amplified data set. The verification experiment shows that the pedestrian detection model based on YOLOv5 has great detection accuracy and recall rate in detecting covered pedestrians.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"39 1","pages":"412-416"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80736323","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9824719
Lin Qiu, Shuo Wang, Jian Wang, Yifei Wang, Wei Huang
Traditional methods of malware detection have difficulty in detecting massive malware variants. Malware detection based on malware visualization has been proved an effective method for identifying unknown malware variants. In order to improve the accuracy and reduce the detection time of above methods, a novel method for malware classification in a light-weight CNN architecture named MalshuffleNet is proposed. The model is customized based on ShuffleNet V2 by adjusting the numbers of the fully connected layer for adopting to malware classification. Empirical results on Malimg dataset indicate that our model achieves 99.03% in accuracy, and identify an unknown malware only taking 5.3 milliseconds on average.
{"title":"Malware Classification based on a Light-weight Architecture of CNN: MalShuffleNet","authors":"Lin Qiu, Shuo Wang, Jian Wang, Yifei Wang, Wei Huang","doi":"10.1109/cvidliccea56201.2022.9824719","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9824719","url":null,"abstract":"Traditional methods of malware detection have difficulty in detecting massive malware variants. Malware detection based on malware visualization has been proved an effective method for identifying unknown malware variants. In order to improve the accuracy and reduce the detection time of above methods, a novel method for malware classification in a light-weight CNN architecture named MalshuffleNet is proposed. The model is customized based on ShuffleNet V2 by adjusting the numbers of the fully connected layer for adopting to malware classification. Empirical results on Malimg dataset indicate that our model achieves 99.03% in accuracy, and identify an unknown malware only taking 5.3 milliseconds on average.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"137 1","pages":"1047-1050"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79740163","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9825366
Daifei Liu, Shengyang Wang
Spheroidity is an important parameter to describe the granulation characteristics of iron ore. Traditionally, physical and chemical analysis methods are used to obtain the spheroidization of iron ore. However, these processes are time-consuming and labor-intensive, and it is difficult to control the accuracy of the results. In this study, image processing and neural networks are used to construct a support vector regression (SVR) iron ore sphericity prediction model from the perspective of information fusion. Three kinds of image texture feature extraction methods are used: Tamura texture feature, gray level co-occurrence matrix (GLCM), and gray level difference statistics (GLDS). Principal component analysis are used to dimensionality reduction of image texture feature parameters. Under the same operating conditions, the results using the SVR model with and without PCA are compared, and the prediction accuracy of these models for iron ore spheroidity are 96.7% and 79.8%, respectively. The results show that the model based on image texture features and PCA-SVR has excellent characteristics, such as fast operating time and high accuracy, for the prediction of iron ore spheroidity, has practical significance in guiding the sintering process of iron ore and can provide further efficient and accurate research on iron ore spheroidity in the future.
{"title":"Prediction of Iron Ore Spheroidity Based on Image Texture Features and PCA-SVR","authors":"Daifei Liu, Shengyang Wang","doi":"10.1109/cvidliccea56201.2022.9825366","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9825366","url":null,"abstract":"Spheroidity is an important parameter to describe the granulation characteristics of iron ore. Traditionally, physical and chemical analysis methods are used to obtain the spheroidization of iron ore. However, these processes are time-consuming and labor-intensive, and it is difficult to control the accuracy of the results. In this study, image processing and neural networks are used to construct a support vector regression (SVR) iron ore sphericity prediction model from the perspective of information fusion. Three kinds of image texture feature extraction methods are used: Tamura texture feature, gray level co-occurrence matrix (GLCM), and gray level difference statistics (GLDS). Principal component analysis are used to dimensionality reduction of image texture feature parameters. Under the same operating conditions, the results using the SVR model with and without PCA are compared, and the prediction accuracy of these models for iron ore spheroidity are 96.7% and 79.8%, respectively. The results show that the model based on image texture features and PCA-SVR has excellent characteristics, such as fast operating time and high accuracy, for the prediction of iron ore spheroidity, has practical significance in guiding the sintering process of iron ore and can provide further efficient and accurate research on iron ore spheroidity in the future.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"115 1","pages":"1-5"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"80868993","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With the continuous development of smart grids, power inspections have become intelligent and sophisticated. This paper proposes a method based on inclined boxes for the automatic position recognition and diagnosis of suspension insulators under a visible light channel. The rotational region convolutional neural networks (R2CNN) algorithm is used to extract the features of large sample images of suspension insulators, and the model is trained to identify and select insulated devices in any direction. The open-source TensorFlow software is used as the identification tool and is combined with related tuning strategies to optimize the model during the training process. The final model’s recognition accuracy was 89.73%. The results prove that this method overcomes the limitations of using axis-aligned boxes for detection, which can provide more accurate position information for diagnoses of suspension insulators. The model has strong robustness in the changing environment, and has certain innovation value and engineering significance.
{"title":"Multi angle location and identification method of suspension insulators based on R2CNN algorithm","authors":"Chao Hou, Yuchen Xing, Ziru Ma, Hai-Fen Liu, Shaotong Pei, Rui Yang, Zhilei Li","doi":"10.1109/cvidliccea56201.2022.9824037","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9824037","url":null,"abstract":"With the continuous development of smart grids, power inspections have become intelligent and sophisticated. This paper proposes a method based on inclined boxes for the automatic position recognition and diagnosis of suspension insulators under a visible light channel. The rotational region convolutional neural networks (R2CNN) algorithm is used to extract the features of large sample images of suspension insulators, and the model is trained to identify and select insulated devices in any direction. The open-source TensorFlow software is used as the identification tool and is combined with related tuning strategies to optimize the model during the training process. The final model’s recognition accuracy was 89.73%. The results prove that this method overcomes the limitations of using axis-aligned boxes for detection, which can provide more accurate position information for diagnoses of suspension insulators. The model has strong robustness in the changing environment, and has certain innovation value and engineering significance.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"27 1","pages":"1213-1216"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87262511","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9823987
Xu Zheng
In this study, the author intends to provide a suitable design for the express distribution path to shorten delivery times. If the route is not well-planned, the delivery time for an express shuttle between cities will be extremely long. The primary purpose of this experimental research is to combine K-means and pointer network optimization to examine the ability of pointer networks in express route planning, improve pointer network performance in the TSP challenge, and obtain a shorter express route planning. The improved K-Pointer-Network is compared to the regular pointer network in this study. According to model theory and experimental data, it can be demonstrated that clustering data samples independently improves computational performance and planning results in many cases, and that when the model is confronted with a large number of test inputs, the K-Pointer-Network outperforms the traditional pointer network and provides relatively good express route planning.
{"title":"K-Pointer-Network for Express Delivery Routes Planning","authors":"Xu Zheng","doi":"10.1109/cvidliccea56201.2022.9823987","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9823987","url":null,"abstract":"In this study, the author intends to provide a suitable design for the express distribution path to shorten delivery times. If the route is not well-planned, the delivery time for an express shuttle between cities will be extremely long. The primary purpose of this experimental research is to combine K-means and pointer network optimization to examine the ability of pointer networks in express route planning, improve pointer network performance in the TSP challenge, and obtain a shorter express route planning. The improved K-Pointer-Network is compared to the regular pointer network in this study. According to model theory and experimental data, it can be demonstrated that clustering data samples independently improves computational performance and planning results in many cases, and that when the model is confronted with a large number of test inputs, the K-Pointer-Network outperforms the traditional pointer network and provides relatively good express route planning.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"257 1","pages":"1193-1197"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79565803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-05-20DOI: 10.1109/cvidliccea56201.2022.9824548
Jiyun Li, Gege Wen, Chen Qian
Brain network plays an important role in the diagnosis of many brain diseases. At present, some related studies are based on the structural or functional connection group of human brain, while others consider the related properties of structural and functional brain networks at the same time. Aiming at the problems of how to dynamically collect richer node interaction information and how to learn more effectively from small samples in the research of brain network fusion, we propose a Random Walk-Grassmann (RW-GM) model to effectively fuse them. Firstly, we obtain the structural connection matrix and the temporal characteristic matrix of the brain from the multi-modal data of each subject. Then, we use random walk algorithm and Grassmann pooling method to integrate the two matrices, in order to integrate the structural connection and the temporal characteristics of the brain, so as to obtain more abundant brain connection information. In order to better carry out small sample learning, we use recursive feature elimination method for feature selection, and put the selected features into support vector machine to get the final classification result. We have carried out four binary classification experiments on ADNI data set, and the classification accuracy is better than that of traditional brain network classification methods.
{"title":"Multi-modal Brain Network Fusion Based on Random Walk-Grassmann Model","authors":"Jiyun Li, Gege Wen, Chen Qian","doi":"10.1109/cvidliccea56201.2022.9824548","DOIUrl":"https://doi.org/10.1109/cvidliccea56201.2022.9824548","url":null,"abstract":"Brain network plays an important role in the diagnosis of many brain diseases. At present, some related studies are based on the structural or functional connection group of human brain, while others consider the related properties of structural and functional brain networks at the same time. Aiming at the problems of how to dynamically collect richer node interaction information and how to learn more effectively from small samples in the research of brain network fusion, we propose a Random Walk-Grassmann (RW-GM) model to effectively fuse them. Firstly, we obtain the structural connection matrix and the temporal characteristic matrix of the brain from the multi-modal data of each subject. Then, we use random walk algorithm and Grassmann pooling method to integrate the two matrices, in order to integrate the structural connection and the temporal characteristics of the brain, so as to obtain more abundant brain connection information. In order to better carry out small sample learning, we use recursive feature elimination method for feature selection, and put the selected features into support vector machine to get the final classification result. We have carried out four binary classification experiments on ADNI data set, and the classification accuracy is better than that of traditional brain network classification methods.","PeriodicalId":23649,"journal":{"name":"Vision","volume":"27 1","pages":"129-134"},"PeriodicalIF":0.0,"publicationDate":"2022-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82041520","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：