Pub Date : 2023-07-02DOI: 10.1109/SSP53291.2023.10208037
B. Unursaikhan, G. Sun, T. Matsui, Gereltuya Amarsanaa
In this paper, we design and develop a vital signs-based mobile medical screening system using cameras (MMSS) to detect possible COVID-19 infection in a non-contact way. The MMSS utilizes different types of cameras, including red-green-blue, depth, and thermal cameras, to measure physiological parameters such as heart rate (HR), respiration rate (RR), and body temperature (BT) in order to detect the infection. We proposed body movement reduction and measurement condition assessment algorithms to acquire reliable physiological signals. Also, we proposed a pixel translation-based computation cost-effective method for setting multiple regions of interest for the cameras’ images. The MMSS-obtained HR, RR, and BT measurement results and the references were correlated significantly with correlation coefficients of 0.97, 0.93, and 0.72, respectively. In clinical testing, the MMSS demonstrated 91% sensitivity and 90% specificity for screening COVID-19 infection.
{"title":"Physiological Parameters-Based Mobile and Non-Contact COVID-19 Screening System Using RGB-Depth-Thermal Cameras","authors":"B. Unursaikhan, G. Sun, T. Matsui, Gereltuya Amarsanaa","doi":"10.1109/SSP53291.2023.10208037","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208037","url":null,"abstract":"In this paper, we design and develop a vital signs-based mobile medical screening system using cameras (MMSS) to detect possible COVID-19 infection in a non-contact way. The MMSS utilizes different types of cameras, including red-green-blue, depth, and thermal cameras, to measure physiological parameters such as heart rate (HR), respiration rate (RR), and body temperature (BT) in order to detect the infection. We proposed body movement reduction and measurement condition assessment algorithms to acquire reliable physiological signals. Also, we proposed a pixel translation-based computation cost-effective method for setting multiple regions of interest for the cameras’ images. The MMSS-obtained HR, RR, and BT measurement results and the references were correlated significantly with correlation coefficients of 0.97, 0.93, and 0.72, respectively. In clinical testing, the MMSS demonstrated 91% sensitivity and 90% specificity for screening COVID-19 infection.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126179499","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10207990
Toan-Van Nguyen, Thien Huynh-The, Vo Nguyen Quoc Bao
This paper studies full-duplex (FD) energy-harvesting Internet-of-Things (IoT) networks, where multiple FD IoT devices are deployed to assist short-packet communications between a source and a robot used in automation factories. Taking into account two residual interference models for FD relays, we propose a full relay selection (FRS) scheme that maximizes the end-to-end signal-to-noise ratio of packet transmissions aiming at improving the block error rate (BLER) and system throughput. Towards real-time settings, we design a deep learning framework based on the FRS scheme to accurately predict the average BLER and throughput via a short inference process. Simulation results show the significant effects of RSI models on the performance of FD IoT networks. Importantly, the DL framework can estimate similar BLER and throughput values as the FRS scheme, but with significantly reduced complexity and execution time, showing the potential of DL design in dealing with complex scenarios of heterogeneous IoT networks.
{"title":"Performance Analysis and Deep Learning Evaluation of URLLC Full-Duplex Energy Harvesting IoT Networks over Nakagami-m Fading Channels","authors":"Toan-Van Nguyen, Thien Huynh-The, Vo Nguyen Quoc Bao","doi":"10.1109/SSP53291.2023.10207990","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10207990","url":null,"abstract":"This paper studies full-duplex (FD) energy-harvesting Internet-of-Things (IoT) networks, where multiple FD IoT devices are deployed to assist short-packet communications between a source and a robot used in automation factories. Taking into account two residual interference models for FD relays, we propose a full relay selection (FRS) scheme that maximizes the end-to-end signal-to-noise ratio of packet transmissions aiming at improving the block error rate (BLER) and system throughput. Towards real-time settings, we design a deep learning framework based on the FRS scheme to accurately predict the average BLER and throughput via a short inference process. Simulation results show the significant effects of RSI models on the performance of FD IoT networks. Importantly, the DL framework can estimate similar BLER and throughput values as the FRS scheme, but with significantly reduced complexity and execution time, showing the potential of DL design in dealing with complex scenarios of heterogeneous IoT networks.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114586901","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208053
Naoki Yoshimura, Toshihisa Tanaka, Yuta Inaba
The problem of estimating imagined music from electroencephalogram (EEG) is very challenging. In this paper, we focused on beats (pulse trains of single notes), one of the components of music, and attempted to estimate imagined beats from an EEG. First, we presented two types of beat patterns and asked 17 experimental participants to imagine them. Next, the imagined beat pulses were estimated from the EEG during the task based on spatiotemporal convolutional neural network models. We employed a CNN and an EEGNet to evaluate the model’s performance with binary cross entropy and focal loss as AUC and F1-measure. Although AUCs between the CNN model and EEGNet are competitive, the number of parameters of the EEGNet is much smaller than that of the CNN. Moreover, we have observed the effect of the loss functions in the F1-measure. Overall, the EEGNet model with the focal loss efficiently performed in imagined beat identification.
{"title":"Estimation of Imagined Rhythms from EEG by Spatiotemporal Convolutional Neural Networks","authors":"Naoki Yoshimura, Toshihisa Tanaka, Yuta Inaba","doi":"10.1109/SSP53291.2023.10208053","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208053","url":null,"abstract":"The problem of estimating imagined music from electroencephalogram (EEG) is very challenging. In this paper, we focused on beats (pulse trains of single notes), one of the components of music, and attempted to estimate imagined beats from an EEG. First, we presented two types of beat patterns and asked 17 experimental participants to imagine them. Next, the imagined beat pulses were estimated from the EEG during the task based on spatiotemporal convolutional neural network models. We employed a CNN and an EEGNet to evaluate the model’s performance with binary cross entropy and focal loss as AUC and F1-measure. Although AUCs between the CNN model and EEGNet are competitive, the number of parameters of the EEGNet is much smaller than that of the CNN. Moreover, we have observed the effect of the loss functions in the F1-measure. Overall, the EEGNet model with the focal loss efficiently performed in imagined beat identification.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129323555","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}
Object counting is the process of determining the count of objects in images using computer vision techniques. In this paper, we employ several state-of-the-art object detection and tracking algorithms to solve the object counting problem in image regions of interest (ROI) on Vietnamese streets. Specifically, we propose video-based methods for counting vehicles in various weather conditions and low-light environments, a new dataset for Vietnamese streets, and retrain the scratch model on the new dataset. A video is processed in three phases, including object detection with YOLO (You Only Look Once), tracking with StrongSORT, and vehicle counting in ROI. The experimental analysis of real-world video footage demonstrates that the proposed method can accurately detect, monitor, and count vehicles. In addition, by using our collected dataset, the proposed method performs significantly better than the pretrained YOLO model.
{"title":"Vehicle Counting on Vietnamese Street","authors":"Khoa Minh Truong, Q. Dinh, Tuan-Duc Nguyen, Thanh Nguyen Nhut","doi":"10.1109/SSP53291.2023.10208075","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208075","url":null,"abstract":"Object counting is the process of determining the count of objects in images using computer vision techniques. In this paper, we employ several state-of-the-art object detection and tracking algorithms to solve the object counting problem in image regions of interest (ROI) on Vietnamese streets. Specifically, we propose video-based methods for counting vehicles in various weather conditions and low-light environments, a new dataset for Vietnamese streets, and retrain the scratch model on the new dataset. A video is processed in three phases, including object detection with YOLO (You Only Look Once), tracking with StrongSORT, and vehicle counting in ROI. The experimental analysis of real-world video footage demonstrates that the proposed method can accurately detect, monitor, and count vehicles. In addition, by using our collected dataset, the proposed method performs significantly better than the pretrained YOLO model.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129691779","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10207963
T. Lee, H. W. Chan, K. Leo, E. Chew, Ling Zhao, Saeid Sanei
Time series data is often processed to extract features which better explain the sources and structure of the data. However, these processes make underlying assumptions about the nature of the time series. Two important intrinsic properties are the linearity and stationarity of the data. The large corpora on time series analyses include domains of economics, physics and engineering – thus cross domain approaches can yield useful insights into the data. Here we look at data from accelerometers, an important class of sensors. We employ widely used time series tests to provide novel analyses to establish their linear and stationary structure. This provides useful insights into the underlying processes which are being sensed and guide the type of temporal features, any preprocessing needed and suitable analyses to be performed. We briefly mention the use of this in a machine learning application.
{"title":"Intrinsic Properties of Human Accelerometer Data for Machine Learning","authors":"T. Lee, H. W. Chan, K. Leo, E. Chew, Ling Zhao, Saeid Sanei","doi":"10.1109/SSP53291.2023.10207963","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10207963","url":null,"abstract":"Time series data is often processed to extract features which better explain the sources and structure of the data. However, these processes make underlying assumptions about the nature of the time series. Two important intrinsic properties are the linearity and stationarity of the data. The large corpora on time series analyses include domains of economics, physics and engineering – thus cross domain approaches can yield useful insights into the data. Here we look at data from accelerometers, an important class of sensors. We employ widely used time series tests to provide novel analyses to establish their linear and stationary structure. This provides useful insights into the underlying processes which are being sensed and guide the type of temporal features, any preprocessing needed and suitable analyses to be performed. We briefly mention the use of this in a machine learning application.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129735806","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}
This paper proposes an extended system for acquiring external sensor data during polysomnography (PSG) tests. The proposed method intends to provide means for integrating external sensor devices with existing PSG systems in the sleep clinic without modification to the original system. The acquired external sensor data is temporally synchronized with the PSG data through a trigger signal and can be used in subsequent applications. Using our proposed system, we acquire ten additional channels via Polar Verity Sense. We validate the acquired data qualitatively through signal visualization. Then, we compare the derived heart rate (HR) from PPG with those from an electrocardiogram (ECG) in the PSG device to validate our system quantitatively. The result reveals a fair error between PPG and ECG HR, demonstrating an acceptable performance. In addition to the PPG data acquisition, the proposed method can be employed on different external sensors to produce various databases for sleep studies.
{"title":"An Extended System for External Sensors Data Acquisition and Validation During Conducting Polysomnography","authors":"Tanut Choksatchawathi, Thitikorn Kaewlee, Guntitat Sawadwuthikul, Busarakum Chaitusaney, N. Jaimchariyatam, Theerawit Wilaiprasitporn, Thapanun Sudhawiyangkul","doi":"10.1109/SSP53291.2023.10208050","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208050","url":null,"abstract":"This paper proposes an extended system for acquiring external sensor data during polysomnography (PSG) tests. The proposed method intends to provide means for integrating external sensor devices with existing PSG systems in the sleep clinic without modification to the original system. The acquired external sensor data is temporally synchronized with the PSG data through a trigger signal and can be used in subsequent applications. Using our proposed system, we acquire ten additional channels via Polar Verity Sense. We validate the acquired data qualitatively through signal visualization. Then, we compare the derived heart rate (HR) from PPG with those from an electrocardiogram (ECG) in the PSG device to validate our system quantitatively. The result reveals a fair error between PPG and ECG HR, demonstrating an acceptable performance. In addition to the PPG data acquisition, the proposed method can be employed on different external sensors to produce various databases for sleep studies.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114796249","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208081
Xunmeng Wu, Zai Yang, Zhiqiang Wei, Zongben Xu
We study the source localization problem for constant modulus (CM) signal using a uniform linear array. Existing results on parameter identifiability show that the maximum number of CM signal sources that can be uniquely localized can exceed the number of sensors, but a practical algorithm is still lacking so far. In this paper, we propose a structured matrix recovery technique (SMART) for CM signal source localization. In particular, the source localization problem is cast as a rank-constrained Hankel-Toeplitz matrix-based feasibility problem, in which signal structures are fully exploited. The alternating direction method of multipliers (ADMM) algorithm is applied to solve the resulting rank-constrained problem and the sources are uniquely retrieved from the numerical solution. Numerical results demonstrate that the proposed SMART can localize more sources than sensors.
{"title":"Source Localization for Constant Modulus Signals Using a Structured Matrix Recovery Technique (SMART)","authors":"Xunmeng Wu, Zai Yang, Zhiqiang Wei, Zongben Xu","doi":"10.1109/SSP53291.2023.10208081","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208081","url":null,"abstract":"We study the source localization problem for constant modulus (CM) signal using a uniform linear array. Existing results on parameter identifiability show that the maximum number of CM signal sources that can be uniquely localized can exceed the number of sensors, but a practical algorithm is still lacking so far. In this paper, we propose a structured matrix recovery technique (SMART) for CM signal source localization. In particular, the source localization problem is cast as a rank-constrained Hankel-Toeplitz matrix-based feasibility problem, in which signal structures are fully exploited. The alternating direction method of multipliers (ADMM) algorithm is applied to solve the resulting rank-constrained problem and the sources are uniquely retrieved from the numerical solution. Numerical results demonstrate that the proposed SMART can localize more sources than sensors.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124451373","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10207989
Binh P. Nguyen, T. Nguyen-Vo, Loc Nguyen, Quang H. Trinh, Chalinor Baliuag, T. Do, S. Rahardja
As a common biological event observed in all living creatures, RNA modification is an essential post-transcriptional factor that regulates the activity, localization, and stability of RNAs. Multiple diseases are associated with RNA modification. N6-methyladenosine (6mA) modification of RNA is one of the most frequent events that affect the translational processes and structural stability of modified transcripts and control transcriptional processes in cell state maintenance and transition. To detect 6mA sites in eukaryotic transcriptomes, a number of computational models were developed as online applications to assist experimental scientists in reducing human effort and budget. However, most of those online web servers are now either outdated or inaccessible. In this study, we propose iR6mA-RNN, an effective computational framework using recurrent neural networks and sequence-embedded features, to predict possible 6mA sites in eukaryotic transcriptomes. When tested on an independent test set, the proposed model achieved an area under the receiver operating characteristic curve of 0.7972 and an area under the precision-recall curve of 0.7785. Our model also outperformed the other two existing methods. Results from another sensitivity analysis confirmed the stability of the model as well.
{"title":"iR6mA-RNN: Identifying N6-Methyladenosine Sites in Eukaryotic Transcriptomes using Recurrent Neural Networks and Sequence-embedded Features","authors":"Binh P. Nguyen, T. Nguyen-Vo, Loc Nguyen, Quang H. Trinh, Chalinor Baliuag, T. Do, S. Rahardja","doi":"10.1109/SSP53291.2023.10207989","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10207989","url":null,"abstract":"As a common biological event observed in all living creatures, RNA modification is an essential post-transcriptional factor that regulates the activity, localization, and stability of RNAs. Multiple diseases are associated with RNA modification. N6-methyladenosine (6mA) modification of RNA is one of the most frequent events that affect the translational processes and structural stability of modified transcripts and control transcriptional processes in cell state maintenance and transition. To detect 6mA sites in eukaryotic transcriptomes, a number of computational models were developed as online applications to assist experimental scientists in reducing human effort and budget. However, most of those online web servers are now either outdated or inaccessible. In this study, we propose iR6mA-RNN, an effective computational framework using recurrent neural networks and sequence-embedded features, to predict possible 6mA sites in eukaryotic transcriptomes. When tested on an independent test set, the proposed model achieved an area under the receiver operating characteristic curve of 0.7972 and an area under the precision-recall curve of 0.7785. Our model also outperformed the other two existing methods. Results from another sensitivity analysis confirmed the stability of the model as well.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131890807","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 : 2023-07-02DOI: 10.1109/SSP53291.2023.10208008
Renwei Huang, Haiyan Wei, Zhenlong Xiao
For real-world graph signals, the relationships between two nodes may not always be symmetric. Hence, a directed graph would be more flexible to characterize such relationships between signals. In this paper, we propose a two-stage algorithm to learn directed graphs from the observed data, i.e., designing the graph frequency components and afterward estimating the graph shift matrix. The graph frequency components are designed to improve the sparsity of graph signals in graph frequency domain, and the estimation of directed shift matrix is thereafter modelled as a convex problem, where the structural constraints of graph signals could be taken into account. Such a directed graph shift matrix would greatly facilitate further processing of the associated graph signals such as sampling and graph filtering in frequency domain since the graph frequency components are specifically designed and the signals over the graph are sparse. Numerical results demonstrate the effectiveness of the proposed method.
{"title":"Learning Directed Graphs From Data Under Structural Constraints","authors":"Renwei Huang, Haiyan Wei, Zhenlong Xiao","doi":"10.1109/SSP53291.2023.10208008","DOIUrl":"https://doi.org/10.1109/SSP53291.2023.10208008","url":null,"abstract":"For real-world graph signals, the relationships between two nodes may not always be symmetric. Hence, a directed graph would be more flexible to characterize such relationships between signals. In this paper, we propose a two-stage algorithm to learn directed graphs from the observed data, i.e., designing the graph frequency components and afterward estimating the graph shift matrix. The graph frequency components are designed to improve the sparsity of graph signals in graph frequency domain, and the estimation of directed shift matrix is thereafter modelled as a convex problem, where the structural constraints of graph signals could be taken into account. Such a directed graph shift matrix would greatly facilitate further processing of the associated graph signals such as sampling and graph filtering in frequency domain since the graph frequency components are specifically designed and the signals over the graph are sparse. Numerical results demonstrate the effectiveness of the proposed method.","PeriodicalId":296346,"journal":{"name":"2023 IEEE Statistical Signal Processing Workshop (SSP)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115423090","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：