Pub Date : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179565
Yi-Hsuan Kao, Chun-Ling Lin
This study adopts a pixel color amplification to increase the characteristics of the fundus image to solve inconsistent images quality problem and adopts EfficientNetV2 architecture of deep convolutional neural network (CNN) to detect Diabetic Retinopathy (DR). The results show that the proposed method can achieve 0.9120/87.16% of quadratic weight kappa and accuracy score and proves the efficacy of the proposed approach in DR classification. Thus, our work can help DR patients to reduce the probability of having lifelong blindness.
{"title":"Detection of Diabetic Retinopathy via Pixel Color Amplification Using EfficientNetV2","authors":"Yi-Hsuan Kao, Chun-Ling Lin","doi":"10.1109/ICASI57738.2023.10179565","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179565","url":null,"abstract":"This study adopts a pixel color amplification to increase the characteristics of the fundus image to solve inconsistent images quality problem and adopts EfficientNetV2 architecture of deep convolutional neural network (CNN) to detect Diabetic Retinopathy (DR). The results show that the proposed method can achieve 0.9120/87.16% of quadratic weight kappa and accuracy score and proves the efficacy of the proposed approach in DR classification. Thus, our work can help DR patients to reduce the probability of having lifelong blindness.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130575673","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179583
Chia-Mei Chen, Sung-Yu Huang, Zheng-Xun Cai, Ya-Hui Ou, Jiunn-Wu Lin
The number of documented supply chain attacks has increased over six times nowadays, and the types of supply chain attacks have diversified. Organizations grant suppliers privileged user accounts to perform their tasks which hold the keys to unlocking internal networks. Privilege escalation is a key step for attackers to penetrate a target system network, which makes privileged user accounts attractive to adversaries. This study employs unsupervised machine learning techniques to profile privileged users’ normal behaviors and develops a risk score function to identify their anomalies. The proposed solution has been evaluated with real data, and the experimental results demonstrate that it could discover the anomalies efficiently.
{"title":"Detecting Supply Chain Attacks with Unsupervised Learning","authors":"Chia-Mei Chen, Sung-Yu Huang, Zheng-Xun Cai, Ya-Hui Ou, Jiunn-Wu Lin","doi":"10.1109/ICASI57738.2023.10179583","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179583","url":null,"abstract":"The number of documented supply chain attacks has increased over six times nowadays, and the types of supply chain attacks have diversified. Organizations grant suppliers privileged user accounts to perform their tasks which hold the keys to unlocking internal networks. Privilege escalation is a key step for attackers to penetrate a target system network, which makes privileged user accounts attractive to adversaries. This study employs unsupervised machine learning techniques to profile privileged users’ normal behaviors and develops a risk score function to identify their anomalies. The proposed solution has been evaluated with real data, and the experimental results demonstrate that it could discover the anomalies efficiently.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130961794","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179535
Wei-Chi Hsu, Yen-Chung Chiang
A wideband amplifier designed in a 180-nm SiGe BiCMOS process technology for cryogenic applications is proposed in this conference paper. By using the common-base stage as the input stage and the emitter follower as the final stage, it can have a wideband input/output matching. The proposed amplifier had been test at room temperature by probing test and in a 4K environment by on-board test and it achieved a peak measured gain of 18.95 dB at room temperature and 25 dB gain at 4K. The measured lowest noise Figure (NF) is between 5.8 to 6.4 dB at room temperature. The proposed circuit draws a 13.52 mW dc-power at room temperature and 12.95 mW at 4K from a 2.0-V supply.
{"title":"A 180-nm SiGe BiCMOS Wideband Amplifier For Cryogenic Applications","authors":"Wei-Chi Hsu, Yen-Chung Chiang","doi":"10.1109/ICASI57738.2023.10179535","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179535","url":null,"abstract":"A wideband amplifier designed in a 180-nm SiGe BiCMOS process technology for cryogenic applications is proposed in this conference paper. By using the common-base stage as the input stage and the emitter follower as the final stage, it can have a wideband input/output matching. The proposed amplifier had been test at room temperature by probing test and in a 4K environment by on-board test and it achieved a peak measured gain of 18.95 dB at room temperature and 25 dB gain at 4K. The measured lowest noise Figure (NF) is between 5.8 to 6.4 dB at room temperature. The proposed circuit draws a 13.52 mW dc-power at room temperature and 12.95 mW at 4K from a 2.0-V supply.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123965737","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179589
Kuo-Guan Wu, Cheng-Feng Lai, Min-Kuan Chang
Point cloud registration involves the task of finding the alignment between pairs of point clouds. Recently, deep learning-based approaches have been shown to achieve more accurate results due to the enhanced feature extraction and correspondence matching mechanisms. Most of the deep learning-based approaches estimate the registration parameters directly without explicitly estimating correspondences. Instead, the correspondences are learned implicitly while networks are trained, and the loss functions used are mostly related to errors of registration parameters. RPMNet, one of the state-of-the-art registration methods, includes an additional component in the loss function to maximize the sum of inlier probabilities in order to prevent the problem of labeling most points as outliers. In this paper, we propose to improve the RPMNet through the modification of the loss function by replacing the sum of inlier probabilities with the sum of maximal matching probabilities, with the purpose of increasing the probability of correct correspondence and thus enhancing the registration performance. Simulation results demonstrate that an order of magnitude improvement can be achieved by the proposed method.
{"title":"On the improvement of RPMNet for deep learning-based point cloud registration using a modified loss function","authors":"Kuo-Guan Wu, Cheng-Feng Lai, Min-Kuan Chang","doi":"10.1109/ICASI57738.2023.10179589","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179589","url":null,"abstract":"Point cloud registration involves the task of finding the alignment between pairs of point clouds. Recently, deep learning-based approaches have been shown to achieve more accurate results due to the enhanced feature extraction and correspondence matching mechanisms. Most of the deep learning-based approaches estimate the registration parameters directly without explicitly estimating correspondences. Instead, the correspondences are learned implicitly while networks are trained, and the loss functions used are mostly related to errors of registration parameters. RPMNet, one of the state-of-the-art registration methods, includes an additional component in the loss function to maximize the sum of inlier probabilities in order to prevent the problem of labeling most points as outliers. In this paper, we propose to improve the RPMNet through the modification of the loss function by replacing the sum of inlier probabilities with the sum of maximal matching probabilities, with the purpose of increasing the probability of correct correspondence and thus enhancing the registration performance. Simulation results demonstrate that an order of magnitude improvement can be achieved by the proposed method.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122979587","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179531
Chun-Chieh Chang, Jun-Ren Chen, S. Ueng
Concealed pipes are widely used to transport water, air, natural gas, and sewage in buildings. In this paper, we propose an algorithm for designing concealed pipes in buildings such that the productivity of construction industry is enhanced. When routing concealed pipes in a building, the walls, floors, and ceilings form the workspace of the piping computation. We treat these slabs as vertices and generate edges between them, based on their connectivity. Consequently, the frame of the building is encoded in a graph, which constitutes the top-level representation of the workspace. Then all the slabs are divided into voxels by using regular grids, and the bottom-level representation of the workspace is created. The pathfinding of a pipe is carried out in two stages. First, the routing process generates a shortest path in the top-level graph to identify the slabs penetrated by the pipe. Then, the routing process creates the detail pipe path by using the voxels of these slabs in the bottom-level representation. Once this pipe has been constructed, the voxels occupied by it are deleted from the workspace and the routing computation is repeated to arrange other pipes.
{"title":"Automatic Pipe Routing in Buildings","authors":"Chun-Chieh Chang, Jun-Ren Chen, S. Ueng","doi":"10.1109/ICASI57738.2023.10179531","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179531","url":null,"abstract":"Concealed pipes are widely used to transport water, air, natural gas, and sewage in buildings. In this paper, we propose an algorithm for designing concealed pipes in buildings such that the productivity of construction industry is enhanced. When routing concealed pipes in a building, the walls, floors, and ceilings form the workspace of the piping computation. We treat these slabs as vertices and generate edges between them, based on their connectivity. Consequently, the frame of the building is encoded in a graph, which constitutes the top-level representation of the workspace. Then all the slabs are divided into voxels by using regular grids, and the bottom-level representation of the workspace is created. The pathfinding of a pipe is carried out in two stages. First, the routing process generates a shortest path in the top-level graph to identify the slabs penetrated by the pipe. Then, the routing process creates the detail pipe path by using the voxels of these slabs in the bottom-level representation. Once this pipe has been constructed, the voxels occupied by it are deleted from the workspace and the routing computation is repeated to arrange other pipes.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"114 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124533675","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 advancement of medical technology, the average life expectancy of human beings has been extended, and the aging of the global population has become an unquestionable fact. How to provide continuous and long-term care for an aging society has become an issue that countries all over the world attach importance to. From preventive medicine that connects hospitals and communities to acute medical, chronic medical and people’s living care, overall health, and care, community preventive care, inpatient medical care, and extended home care at home, not only preventive health care at the front end but also reduce and delay the disability, and have complete palliative care treatment at the back end, so that the aging population can receive complete care, and it is an important issue facing the aging population. This paper proposesc a new type of information technology medical care service. Through the remote medical information service platform developed by us, combining the medical resources of the community and the hospital, we develop a cloud information system for data collection and remote consultation and assist the service information needed in the care process through the provision and reminder of mobile information. In the remote medical information service platform, the server, and the case can be more suitable for medical care. Finally, experimental results show that our platform can not only improve medical quality, high-efficiency medical services, but also reduce medical costs, and make up for the shortage of manpower.
{"title":"Medical Integrated Development Platform Based on Community and Hospital","authors":"Jhe-Wei Lin, Cheng-Yan Siao, Ting-Hsuan Chien, Rong-Guey Chang, Mei-Ling Hsu","doi":"10.1109/ICASI57738.2023.10179511","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179511","url":null,"abstract":"With the advancement of medical technology, the average life expectancy of human beings has been extended, and the aging of the global population has become an unquestionable fact. How to provide continuous and long-term care for an aging society has become an issue that countries all over the world attach importance to. From preventive medicine that connects hospitals and communities to acute medical, chronic medical and people’s living care, overall health, and care, community preventive care, inpatient medical care, and extended home care at home, not only preventive health care at the front end but also reduce and delay the disability, and have complete palliative care treatment at the back end, so that the aging population can receive complete care, and it is an important issue facing the aging population. This paper proposesc a new type of information technology medical care service. Through the remote medical information service platform developed by us, combining the medical resources of the community and the hospital, we develop a cloud information system for data collection and remote consultation and assist the service information needed in the care process through the provision and reminder of mobile information. In the remote medical information service platform, the server, and the case can be more suitable for medical care. Finally, experimental results show that our platform can not only improve medical quality, high-efficiency medical services, but also reduce medical costs, and make up for the shortage of manpower.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132606774","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179521
Cheng-Ta Lo, Yi-Hsuan Lee, Jun-Hong Peng
Twitter users usually care about the popularity of their tweets, and retweet count is always a good measure. Along with Neural Networks have achieved outstanding accomplishments, Graph Neural Network (GNN) becomes a new research field. In this article, we select a group of active users on a Twitter page. After observing their recent retweet behaviors, different GNN models are constructed and labeled. These GNN models are used to predict the user retweet behavior and estimate the retweet count in the early stage.
{"title":"Real-time Retweet Count Prediction using GNN Model","authors":"Cheng-Ta Lo, Yi-Hsuan Lee, Jun-Hong Peng","doi":"10.1109/ICASI57738.2023.10179521","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179521","url":null,"abstract":"Twitter users usually care about the popularity of their tweets, and retweet count is always a good measure. Along with Neural Networks have achieved outstanding accomplishments, Graph Neural Network (GNN) becomes a new research field. In this article, we select a group of active users on a Twitter page. After observing their recent retweet behaviors, different GNN models are constructed and labeled. These GNN models are used to predict the user retweet behavior and estimate the retweet count in the early stage.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132301531","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179506
Che-Wei Chang, Kai-Jie Zhang
It is a promising approach to include Non-Volatile Memory (NVM) and DRAM together as the main memory to provide huge memory space for large objects and guarantee short access latency for frequent writes and reads. To provide feasible and suitable memory allocation for all applications on a sever, our solution analyzes applications to derive the read-write ratio, memory footprint, and access pattern of each application. Heterogenous memory allocation scheme is then developed to reduce the makespan of batch jobs.
{"title":"Heterogeneous Memory Allocation Scheme for Batch Jobs with Intel Optane DC Persistent Memory and DRAM","authors":"Che-Wei Chang, Kai-Jie Zhang","doi":"10.1109/ICASI57738.2023.10179506","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179506","url":null,"abstract":"It is a promising approach to include Non-Volatile Memory (NVM) and DRAM together as the main memory to provide huge memory space for large objects and guarantee short access latency for frequent writes and reads. To provide feasible and suitable memory allocation for all applications on a sever, our solution analyzes applications to derive the read-write ratio, memory footprint, and access pattern of each application. Heterogenous memory allocation scheme is then developed to reduce the makespan of batch jobs.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"83 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115734161","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179594
Ching Pan, Fei Lee, Chun-Fa Chang
The Monte Carlo rendering method requires sampling an additional spectral domain to produce advanced light phenomena like chromatic dispersion. However, this extension can lead to chromatic noise and slow convergence, which can be partially addressed through proper sample distribution. In this paper, we introduce a multiple wavelength sampling method that focuses on refractive dispersion effects for a single path. By combining the idea of line sampling used in direct illumination with the specular chain’s property, we find a continuous solution space that we call the spectral interval. We then perform an analytical integration for incoming spectral radiance within this interval. This allows us to achieve low variance and reduced noise at a lower cost compared to naive sampling methods. Finally, our method can be integrated into existing Monte Carlo methods built within path space.
{"title":"Refraction Rendering with Spectral Interval Sampling","authors":"Ching Pan, Fei Lee, Chun-Fa Chang","doi":"10.1109/ICASI57738.2023.10179594","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179594","url":null,"abstract":"The Monte Carlo rendering method requires sampling an additional spectral domain to produce advanced light phenomena like chromatic dispersion. However, this extension can lead to chromatic noise and slow convergence, which can be partially addressed through proper sample distribution. In this paper, we introduce a multiple wavelength sampling method that focuses on refractive dispersion effects for a single path. By combining the idea of line sampling used in direct illumination with the specular chain’s property, we find a continuous solution space that we call the spectral interval. We then perform an analytical integration for incoming spectral radiance within this interval. This allows us to achieve low variance and reduced noise at a lower cost compared to naive sampling methods. Finally, our method can be integrated into existing Monte Carlo methods built within path space.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116269568","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 : 2023-04-21DOI: 10.1109/ICASI57738.2023.10179566
Chung L. Chang, Hung-Wen Chen
This study proposed a deep learning-based approach to recognize various types of objects in images and generate optimal straight-line segments for mobile robots to perform heading corrections in complex environments. Object detection, based on a circular convolutional network framework, was utilized to identify various objects, such as watering strips, lettuce crops, or field furrows, in both the upper and lower regions of the image. Following the processing of multiple images, the center points of objects belonging to the same category were extracted, and a regression analysis method was used to generate a straight line. The slopes of these line segments are estimated, and the average value is calculated alïer determining the heading angle with the vertical line segment in the image through trigonometric operation. The flexibility and robustness of the straight-line detection system are enhanced by using the proposed approach.
{"title":"Straight-line Generation Approach using Deep Learning for Mobile Robot Guidance in Lettuce Fields","authors":"Chung L. Chang, Hung-Wen Chen","doi":"10.1109/ICASI57738.2023.10179566","DOIUrl":"https://doi.org/10.1109/ICASI57738.2023.10179566","url":null,"abstract":"This study proposed a deep learning-based approach to recognize various types of objects in images and generate optimal straight-line segments for mobile robots to perform heading corrections in complex environments. Object detection, based on a circular convolutional network framework, was utilized to identify various objects, such as watering strips, lettuce crops, or field furrows, in both the upper and lower regions of the image. Following the processing of multiple images, the center points of objects belonging to the same category were extracted, and a regression analysis method was used to generate a straight line. The slopes of these line segments are estimated, and the average value is calculated alïer determining the heading angle with the vertical line segment in the image through trigonometric operation. The flexibility and robustness of the straight-line detection system are enhanced by using the proposed approach.","PeriodicalId":281254,"journal":{"name":"2023 9th International Conference on Applied System Innovation (ICASI)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114654069","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：