Pub Date : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135238
Jin Xiao, Runhua Wang
Deep neural networks are widely used in speech recognition and face verification with excellent performance, and they are gradually applied and developed in the field of customer credit scoring. Traditional credit scoring work relies on the two-step modeling process of feature processing and model building, which cannot effectively balance data dimensionality and model performance. Based on this, we put forward a triplet deep neural network model for customer credit scoring. This model makes use of the feature that deep neural networks and metric learning can efficiently extract and utilize data feature information so that two samples with the same label are embedded tightly while two samples with different labels are embedded loosely, so as to improve the accuracy of credit scoring. All experiments are conducted on three customer credit scoring datasets. We select accuracy, precision, recall, f1-score and AUC to evaluate the classification performance of all models. The experiments show that the triplet deep neural networks model can perform customer credit scoring more accurately compared with the now commonly used random forest (RF), deep neural networks (DNN), logistic regression (LR), k-nearest neighbor (KNN) and support vector machine (SVM).
{"title":"A Triplet Deep Neural Networks Model for Customer Credit Scoring","authors":"Jin Xiao, Runhua Wang","doi":"10.1109/ICCECE58074.2023.10135238","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135238","url":null,"abstract":"Deep neural networks are widely used in speech recognition and face verification with excellent performance, and they are gradually applied and developed in the field of customer credit scoring. Traditional credit scoring work relies on the two-step modeling process of feature processing and model building, which cannot effectively balance data dimensionality and model performance. Based on this, we put forward a triplet deep neural network model for customer credit scoring. This model makes use of the feature that deep neural networks and metric learning can efficiently extract and utilize data feature information so that two samples with the same label are embedded tightly while two samples with different labels are embedded loosely, so as to improve the accuracy of credit scoring. All experiments are conducted on three customer credit scoring datasets. We select accuracy, precision, recall, f1-score and AUC to evaluate the classification performance of all models. The experiments show that the triplet deep neural networks model can perform customer credit scoring more accurately compared with the now commonly used random forest (RF), deep neural networks (DNN), logistic regression (LR), k-nearest neighbor (KNN) and support vector machine (SVM).","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128973514","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}
Self-supervised learning typically suffers from lacking contrastive pairs and extracting unrepresentative vectors. To handle above mentioned challenges, this paper introduces a novel self-supervised learning framework that integrates the location-based sampling manner and a well-designed dimensionality reduction module. In the location-based sampling module, this paper embeds a multi-crop sampling paradigm into the memory bank-based framework. In the dimensionality reduction module, this paper introduces a principal component dimensionality reduction to capture the most comprehensive features. Experiments on popular datasets demonstrate the superior performance of our proposed method.
{"title":"Improved contrastive learning with MoCo framework","authors":"Yihan Li, Qingmin Liu, Ling Zhou, Wenyi Zhao, Y. Tian, Weidong Zhang","doi":"10.1109/ICCECE58074.2023.10135455","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135455","url":null,"abstract":"Self-supervised learning typically suffers from lacking contrastive pairs and extracting unrepresentative vectors. To handle above mentioned challenges, this paper introduces a novel self-supervised learning framework that integrates the location-based sampling manner and a well-designed dimensionality reduction module. In the location-based sampling module, this paper embeds a multi-crop sampling paradigm into the memory bank-based framework. In the dimensionality reduction module, this paper introduces a principal component dimensionality reduction to capture the most comprehensive features. Experiments on popular datasets demonstrate the superior performance of our proposed method.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129311593","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135349
Teng Wang
Android applications (a.k.a., Android apps) have developed rapidly in the last decade, and have become an indispensable part in people's lives. However, it is challenging to guarantee their quality and reliability. The prevalence and severity of Android apps issues have driven the design and development of a number of detection and testing techniques. However, these techniques mainly target the generation of test sequences using GUI events in the applications, lacking of attentions to complex Android system configurations. In this paper, we conducted an in-depth study on real-world Android bugs related with configurations from 20 open-source popular Android applications, to help understand the characteristics of these bugs. We find the majority of configurations-related Android bugs would lead to catastrophic consequences, e.g., crash and hang. Based on the study, we design and implement ConfigDroid, a tool for configuration-aware GUI testing of Android applications. We use 10 open-source popular Android applications to evaluate the effectiveness. The result shows that, ConfigDroid can detect 4 more unique configuration-related crashes than state-of-the-art tools, Monkey and Stoat.
{"title":"ConfigDroid: Configuration-aware GUI testing of Android Applications","authors":"Teng Wang","doi":"10.1109/ICCECE58074.2023.10135349","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135349","url":null,"abstract":"Android applications (a.k.a., Android apps) have developed rapidly in the last decade, and have become an indispensable part in people's lives. However, it is challenging to guarantee their quality and reliability. The prevalence and severity of Android apps issues have driven the design and development of a number of detection and testing techniques. However, these techniques mainly target the generation of test sequences using GUI events in the applications, lacking of attentions to complex Android system configurations. In this paper, we conducted an in-depth study on real-world Android bugs related with configurations from 20 open-source popular Android applications, to help understand the characteristics of these bugs. We find the majority of configurations-related Android bugs would lead to catastrophic consequences, e.g., crash and hang. Based on the study, we design and implement ConfigDroid, a tool for configuration-aware GUI testing of Android applications. We use 10 open-source popular Android applications to evaluate the effectiveness. The result shows that, ConfigDroid can detect 4 more unique configuration-related crashes than state-of-the-art tools, Monkey and Stoat.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128623851","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135506
Yangyang Cao, Songtao Gao, Yiming Yu, Xiangchen Ma
Taking 5G as the starting point, the intelligent coal mine system formed by deep integration with artificial intelligence, big data, intelligent robots and other technologies can improve the safety production capacity of the mining area, strengthen the operation and maintenance management capacity, and promote the development of traditional coal mines towards unmanned, visual, automatic and intelligent trends. The actual production applications in the coal industry mainly include six business scenarios. The network requirements can be summarized into two categories: large bandwidth and low delay. The 5G RAN solution in the coal industry needs to be selected based on the coal mine type, and further superimpose the private network technology scheme according to the differentiated business requirements.
{"title":"Research on 5G Radio Access Network(RAN) Solution for Coal Mine Industry","authors":"Yangyang Cao, Songtao Gao, Yiming Yu, Xiangchen Ma","doi":"10.1109/ICCECE58074.2023.10135506","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135506","url":null,"abstract":"Taking 5G as the starting point, the intelligent coal mine system formed by deep integration with artificial intelligence, big data, intelligent robots and other technologies can improve the safety production capacity of the mining area, strengthen the operation and maintenance management capacity, and promote the development of traditional coal mines towards unmanned, visual, automatic and intelligent trends. The actual production applications in the coal industry mainly include six business scenarios. The network requirements can be summarized into two categories: large bandwidth and low delay. The 5G RAN solution in the coal industry needs to be selected based on the coal mine type, and further superimpose the private network technology scheme according to the differentiated business requirements.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115523126","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135502
Jiaming Xue, Shanlin Sun, Haimeng Zhao, Wei Chen
With the aim of providing a reliable prediction model for vegetation detection and ground classification, a multispectral dataset was produced for semantic segmentation, which utilizes multispectral UAV images and is based on a combination of support vector machines and manual annotation. Also, a 3D-UNet model is proposed on which the dataset is trained and experiments show that the model has achieved 89.9 % prediction for the validation set.
{"title":"Vegetation Classification of UAV Multispectral Remote Sensing Images Based on Deep Learning","authors":"Jiaming Xue, Shanlin Sun, Haimeng Zhao, Wei Chen","doi":"10.1109/ICCECE58074.2023.10135502","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135502","url":null,"abstract":"With the aim of providing a reliable prediction model for vegetation detection and ground classification, a multispectral dataset was produced for semantic segmentation, which utilizes multispectral UAV images and is based on a combination of support vector machines and manual annotation. Also, a 3D-UNet model is proposed on which the dataset is trained and experiments show that the model has achieved 89.9 % prediction for the validation set.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114780980","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135365
Zi-ang Chen, Xing Zhang, Liang Wang, Yunfei Xia
The path planning algorithm for moving objects has high complexity, and the automatic path planning ability is poor, which cannot deal with complex practical environmental problems. A fast path planning algorithm based on RRT-Star is proposed. First, for the diagonal obstacles in path planning, the deadlock back-off method is used to realize obstacle detection, which effectively improves the safety of the path. Second, as it progresses, the algorithm uses a step size adjustment function to expand the step size, thereby increasing the speed at which the random tree can explore this space. In addition, based on the RRT-Star algorithm, the target deviation strategy is introduced, and the initial pheromone allocation principle is proposed. Finally, the pheromone is classified, and the pheromone on each path is superimposed according to the optimization objective. The results show that the RRT-Star fast path planning efficiency and the number of iterations are significantly better than the RRT algorithm and the ant colony algorithm.
{"title":"A Fast Path Planning Method Based on RRT Star Algorithm","authors":"Zi-ang Chen, Xing Zhang, Liang Wang, Yunfei Xia","doi":"10.1109/ICCECE58074.2023.10135365","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135365","url":null,"abstract":"The path planning algorithm for moving objects has high complexity, and the automatic path planning ability is poor, which cannot deal with complex practical environmental problems. A fast path planning algorithm based on RRT-Star is proposed. First, for the diagonal obstacles in path planning, the deadlock back-off method is used to realize obstacle detection, which effectively improves the safety of the path. Second, as it progresses, the algorithm uses a step size adjustment function to expand the step size, thereby increasing the speed at which the random tree can explore this space. In addition, based on the RRT-Star algorithm, the target deviation strategy is introduced, and the initial pheromone allocation principle is proposed. Finally, the pheromone is classified, and the pheromone on each path is superimposed according to the optimization objective. The results show that the RRT-Star fast path planning efficiency and the number of iterations are significantly better than the RRT algorithm and the ant colony algorithm.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127221001","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135301
Guangxiao Niu
Image denoising (DN), demosaicing (DM) and super-resolution (SR) are the key tasks of the low-level vision. Joint demosaicing, denoising and Super-resolution (JDDSR) can effectively improve the image quality. However, the previous methods explored the feasibility of tasks more than the characteristics of DM, DN and SR. Meanwhile, the joint training also brought computational burden and the three tasks process information at different frequencies. DN and DM pay more attention to low-frequency (LF) information, while SR is used to recover the lost high-frequency (HF) information. In this work, we use the way of Laplace pyramid to separate the HF and LF of the image, and use different branches to learn the information of different frequencies. In order to reduce the computational burden, we redesign the network architecture and use the form of non-parametric up-sampling to generate the results. Experiments demonstrate that our method can achieve results similar to existing methods with very small computational effort and storage.
{"title":"Frequency Decomposition Network for Fast Joint Image Demosaic, Denoising and Super-Resolution","authors":"Guangxiao Niu","doi":"10.1109/ICCECE58074.2023.10135301","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135301","url":null,"abstract":"Image denoising (DN), demosaicing (DM) and super-resolution (SR) are the key tasks of the low-level vision. Joint demosaicing, denoising and Super-resolution (JDDSR) can effectively improve the image quality. However, the previous methods explored the feasibility of tasks more than the characteristics of DM, DN and SR. Meanwhile, the joint training also brought computational burden and the three tasks process information at different frequencies. DN and DM pay more attention to low-frequency (LF) information, while SR is used to recover the lost high-frequency (HF) information. In this work, we use the way of Laplace pyramid to separate the HF and LF of the image, and use different branches to learn the information of different frequencies. In order to reduce the computational burden, we redesign the network architecture and use the form of non-parametric up-sampling to generate the results. Experiments demonstrate that our method can achieve results similar to existing methods with very small computational effort and storage.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127010471","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135293
Junqi Zhou, Yanfeng Li, Houjin Chen
Robust local feature detection and description are crucial in a huge of computer vision tasks. However, current feature descriptors lack geometric information and spatial features. In this paper, we propose a novel spatial and geometric feature fusion architecture, which can effectively tackle these problems. The proposed architecture includes three aspects. 1) A multiple low-level feature fusion (ML2F2) subnetworks, which could improve the ability to extract geometric information through the weighted fusion features. 2) A subnetwork for gradient feature extraction (GFE) which could validly extract the spatial features by encoding the horizontal and vertical gradients of an image. 3) Effective spatial geometric feature fusion (SGFF) module to alternatively fuse the above subnetworks. Experiments on the task of Image Matching and Long-Term Visual Localization show that the proposed method is superior to most advanced local feature descriptors.
{"title":"Learning Spatial and Geometric Information for Robust Features","authors":"Junqi Zhou, Yanfeng Li, Houjin Chen","doi":"10.1109/ICCECE58074.2023.10135293","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135293","url":null,"abstract":"Robust local feature detection and description are crucial in a huge of computer vision tasks. However, current feature descriptors lack geometric information and spatial features. In this paper, we propose a novel spatial and geometric feature fusion architecture, which can effectively tackle these problems. The proposed architecture includes three aspects. 1) A multiple low-level feature fusion (ML2F2) subnetworks, which could improve the ability to extract geometric information through the weighted fusion features. 2) A subnetwork for gradient feature extraction (GFE) which could validly extract the spatial features by encoding the horizontal and vertical gradients of an image. 3) Effective spatial geometric feature fusion (SGFF) module to alternatively fuse the above subnetworks. Experiments on the task of Image Matching and Long-Term Visual Localization show that the proposed method is superior to most advanced local feature descriptors.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127612826","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135295
Minghe Xia, Wenhao Liu, Dibin Zhou
In a surgical navigation system, point cloud registration technology determines the accuracy of surgical navigation. Intelligent analysis of artificial joints to achieve higher quality measurement has become an urgent problem in point cloud registration. Based on deep learning, the experimental design and registration test method of point cloud acquisition based on artificial joint replacement are proposed. Experimental results show that the new algorithm can improve surgical accuracy. (Abstract)
{"title":"Research on test method of point cloud registration based on joint replacement","authors":"Minghe Xia, Wenhao Liu, Dibin Zhou","doi":"10.1109/ICCECE58074.2023.10135295","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135295","url":null,"abstract":"In a surgical navigation system, point cloud registration technology determines the accuracy of surgical navigation. Intelligent analysis of artificial joints to achieve higher quality measurement has become an urgent problem in point cloud registration. Based on deep learning, the experimental design and registration test method of point cloud acquisition based on artificial joint replacement are proposed. Experimental results show that the new algorithm can improve surgical accuracy. (Abstract)","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125979132","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 : 2023-01-06DOI: 10.1109/ICCECE58074.2023.10135303
Xin Wang, Boxuan Zhang, Ya'nan Li
With the promotion of the “new four automobile modernizations” and the rise of users' awareness of travel service demand, user experience has penetrated into the whole process from R & D (research and development) to sales of automotive products. Based on the questionnaire survey data, this paper uses K-means algorithm to subdivide new energy users. Firstly, factor analysis and principal component analysis are used to analyze users' values and career level, then K-means clustering is carried out on this basis, and user characteristics are visually analyzed. Finally, new energy users are divided into six categories, and the car purchase preferences of each category of users are deeply analyzed, which has important theoretical and practical significance for enterprises to accurately grasp users' needs and clarify the future research and development direction.
{"title":"Research on New Energy User Characteristics Based on Machine Learning Algorithm","authors":"Xin Wang, Boxuan Zhang, Ya'nan Li","doi":"10.1109/ICCECE58074.2023.10135303","DOIUrl":"https://doi.org/10.1109/ICCECE58074.2023.10135303","url":null,"abstract":"With the promotion of the “new four automobile modernizations” and the rise of users' awareness of travel service demand, user experience has penetrated into the whole process from R & D (research and development) to sales of automotive products. Based on the questionnaire survey data, this paper uses K-means algorithm to subdivide new energy users. Firstly, factor analysis and principal component analysis are used to analyze users' values and career level, then K-means clustering is carried out on this basis, and user characteristics are visually analyzed. Finally, new energy users are divided into six categories, and the car purchase preferences of each category of users are deeply analyzed, which has important theoretical and practical significance for enterprises to accurately grasp users' needs and clarify the future research and development direction.","PeriodicalId":120030,"journal":{"name":"2023 3rd International Conference on Consumer Electronics and Computer Engineering (ICCECE)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126582084","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：