Pub Date : 2001-07-08DOI: 10.1109/EVENT.2001.938863
Anjum Ali, J. Aggarwal
This paper presents a methodology for automatic segmentation and recognition of continuous human activity. We segment a continuous human activity into separate actions and correctly identify each action. The camera views the subject from the lateral view: there are no distinct breaks or pauses between the execution of different actions. We have no prior knowledge about the commencement or termination of each action. We compute the angles subtended by three major components of the body with the vertical axis, namely the torso, the upper component of the leg and the lower component of the leg. Using these three angles as a feature vector we classify frames into breakpoint and non-breakpoint frames. Breakpoints indicate an action's commencement or termination. We use single action sequences for the training data set. The test sequences, on the other hand are continuous sequences of human activity that consist of three or more actions in succession. The system has been tested on continuous activity sequences containing actions such as walking, sitting down, standing up, bending, getting up, squatting and rising. It detects the breakpoints and classifies the actions between them.
{"title":"Segmentation and recognition of continuous human activity","authors":"Anjum Ali, J. Aggarwal","doi":"10.1109/EVENT.2001.938863","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938863","url":null,"abstract":"This paper presents a methodology for automatic segmentation and recognition of continuous human activity. We segment a continuous human activity into separate actions and correctly identify each action. The camera views the subject from the lateral view: there are no distinct breaks or pauses between the execution of different actions. We have no prior knowledge about the commencement or termination of each action. We compute the angles subtended by three major components of the body with the vertical axis, namely the torso, the upper component of the leg and the lower component of the leg. Using these three angles as a feature vector we classify frames into breakpoint and non-breakpoint frames. Breakpoints indicate an action's commencement or termination. We use single action sequences for the training data set. The test sequences, on the other hand are continuous sequences of human activity that consist of three or more actions in succession. The system has been tested on continuous activity sequences containing actions such as walking, sitting down, standing up, bending, getting up, squatting and rising. It detects the breakpoints and classifies the actions between them.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115748044","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938868
T. Syeda-Mahmood, M. Alex O. Vasilescu, Saratendu Sethi
A first step towards an understanding of the semantic content in a video is the reliable detection and recognition of actions performed by objects. This is a difficult problem due to the enormous variability in an action's appearance when seen from different viewpoints and/or at different times. In this paper we address the recognition of actions by taking a novel approach that models actions as special types of 3D objects. Specifically, we observe that any action can be represented as a generalized cylinder, called the action cylinder. Reliable recognition is achieved by recovering the viewpoint transformation between the reference (model) and given action cylinders. A set of 8 corresponding points from time-wise corresponding cross-sections is shown to be sufficient to align the two cylinders under perspective projection. A surprising conclusion from visualizing actions as objects is that rigid, articulated, and nonrigid actions can all be modeled in a uniform framework.
{"title":"Recognizing action events from multiple viewpoints","authors":"T. Syeda-Mahmood, M. Alex O. Vasilescu, Saratendu Sethi","doi":"10.1109/EVENT.2001.938868","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938868","url":null,"abstract":"A first step towards an understanding of the semantic content in a video is the reliable detection and recognition of actions performed by objects. This is a difficult problem due to the enormous variability in an action's appearance when seen from different viewpoints and/or at different times. In this paper we address the recognition of actions by taking a novel approach that models actions as special types of 3D objects. Specifically, we observe that any action can be represented as a generalized cylinder, called the action cylinder. Reliable recognition is achieved by recovering the viewpoint transformation between the reference (model) and given action cylinders. A set of 8 corresponding points from time-wise corresponding cross-sections is shown to be sufficient to align the two cylinders under perspective projection. A surprising conclusion from visualizing actions as objects is that rigid, articulated, and nonrigid actions can all be modeled in a uniform framework.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"46 12","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132286511","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938870
M. Slaney, D. Ponceleón, James Kaufman
This paper describes a new representation for the audio and visual information in a video signal. We use reduce the dimensionality of the signals with singular-value decomposition (SVD) or mel-frequency cepstral coefficients (MFCC). We apply these transforms to word, (word transcript, semantic space or latent semantic indexing), image (color histogram data) and audio (timbre) data. Using scale-space techniques we find large jumps in a video's path, which are evidence for events. We use these techniques to analyze the temporal properties of the audio and image data in a video. This analysis creates a hierarchical segmentation of the video, or a table-of-contents, from both audio and the image data.
{"title":"Temporal events in all dimensions and scales","authors":"M. Slaney, D. Ponceleón, James Kaufman","doi":"10.1109/EVENT.2001.938870","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938870","url":null,"abstract":"This paper describes a new representation for the audio and visual information in a video signal. We use reduce the dimensionality of the signals with singular-value decomposition (SVD) or mel-frequency cepstral coefficients (MFCC). We apply these transforms to word, (word transcript, semantic space or latent semantic indexing), image (color histogram data) and audio (timbre) data. Using scale-space techniques we find large jumps in a video's path, which are evidence for events. We use these techniques to analyze the temporal properties of the audio and image data in a video. This analysis creates a hierarchical segmentation of the video, or a table-of-contents, from both audio and the image data.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134312243","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938865
N. Krahnstoever, M. Yeasin, Rajeev Sharma
We propose a framework for detecting, tracking and analyzing non-rigid motion based on learned motion patterns. The framework features an appearance based approach to represent the spatial information and hidden Markov models (HMM) to encode the temporal dynamics of the time varying visual patterns. The low level spatial feature extraction is fused with the temporal analysis, providing a unified spatio-temporal approach to common detection, tracking and classification problems. This is a promising approach for many classes of human motion patterns. Visual tracking is achieved by extracting the most probable sequence of target locations from a video stream using a combination of random sampling and the forward procedure from HMM theory. The method allows us to perform a set of important tasks such as activity recognition, gait-analysis and keyframe extraction. The efficacy of the method is shown on both natural and synthetic test sequences.
{"title":"Towards a unified framework for tracking and analysis of human motion","authors":"N. Krahnstoever, M. Yeasin, Rajeev Sharma","doi":"10.1109/EVENT.2001.938865","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938865","url":null,"abstract":"We propose a framework for detecting, tracking and analyzing non-rigid motion based on learned motion patterns. The framework features an appearance based approach to represent the spatial information and hidden Markov models (HMM) to encode the temporal dynamics of the time varying visual patterns. The low level spatial feature extraction is fused with the temporal analysis, providing a unified spatio-temporal approach to common detection, tracking and classification problems. This is a promising approach for many classes of human motion patterns. Visual tracking is achieved by extracting the most probable sequence of target locations from a video stream using a combination of random sampling and the forward procedure from HMM theory. The method allows us to perform a set of important tasks such as activity recognition, gait-analysis and keyframe extraction. The efficacy of the method is shown on both natural and synthetic test sequences.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127569892","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938872
J. Hoey
We consider the problem of unsupervised classification of temporal sequences of facial expressions in video. This problem arises in the design of an adaptive visual agent, which must be capable of identifying appropriate classes of visual events without supervision to effectively complete its tasks. We present a multilevel dynamic Bayesian network that learns the high-level dynamics of facial expressions simultaneously, with models of the expressions themselves. We show how the parameters of the model can be learned in a scalable and efficient way. We present preliminary results using real video data and a class of simulated dynamic event models. The results show that our model correctly classifies the input data comparably to a standard event classification approach, while also learning the high-level model parameters.
{"title":"Hierarchical unsupervised learning of facial expression categories","authors":"J. Hoey","doi":"10.1109/EVENT.2001.938872","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938872","url":null,"abstract":"We consider the problem of unsupervised classification of temporal sequences of facial expressions in video. This problem arises in the design of an adaptive visual agent, which must be capable of identifying appropriate classes of visual events without supervision to effectively complete its tasks. We present a multilevel dynamic Bayesian network that learns the high-level dynamics of facial expressions simultaneously, with models of the expressions themselves. We show how the parameters of the model can be learned in a scalable and efficient way. We present preliminary results using real video data and a class of simulated dynamic event models. The results show that our model correctly classifies the input data comparably to a standard event classification approach, while also learning the high-level model parameters.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115794398","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938861
J. Burns
In airborne video, objects are tracked from a moving camera and often imaged at very low resolution. The camera movement makes it difficult to determine whether or not an object is in motion; the low-resolution imagery makes it difficult to classify the objects and their activities. When comparable, the object's georeferenced trajectory contains useful information for the solution of both of these problems. We describe a novel technique for detecting independent movement by analyzing georeferenced object motion relative to the trajectory of the camera. The method is demonstrated on over a hundred objects and parallax artifacts, and its performance is analyzed relative to difficult object behavior and camera model errors. We also describe a new method for classifying objects and events using features of georeferenced trajectories, such as duration of acceleration, measured at key phases of the events. These features, combined with the periodicity of the image motion, are successfully used classify events in the domain of person-vehicle interactions.
{"title":"Detecting independently moving objects and their interactions in georeferenced airborne video","authors":"J. Burns","doi":"10.1109/EVENT.2001.938861","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938861","url":null,"abstract":"In airborne video, objects are tracked from a moving camera and often imaged at very low resolution. The camera movement makes it difficult to determine whether or not an object is in motion; the low-resolution imagery makes it difficult to classify the objects and their activities. When comparable, the object's georeferenced trajectory contains useful information for the solution of both of these problems. We describe a novel technique for detecting independent movement by analyzing georeferenced object motion relative to the trajectory of the camera. The method is demonstrated on over a hundred objects and parallax artifacts, and its performance is analyzed relative to difficult object behavior and camera model errors. We also describe a new method for classifying objects and events using features of georeferenced trajectories, such as duration of acceleration, measured at key phases of the events. These features, combined with the periodicity of the image motion, are successfully used classify events in the domain of person-vehicle interactions.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125347836","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938864
James W. Davis
There has been increasing interest in computer analysis and recognition of human motion. Previously we presented an efficient real-time approach for representing human motion using a compact "motion history image" (MHI). Recognition was achieved by statistically matching moment-based features. To address previous problems related to global analysis and limited recognition, we present a hierarchical extension to the original MHI framework to compute dense (local) motion flow directly from the MHI. A hierarchical partitioning of motions by speed in an MHI pyramid enables efficient calculation of image motions using fixed-size gradient operators. To characterize the resulting motion field, a polar histogram of motion orientations is described. The hierarchical MHI approach remains a computationally inexpensive method for analysis of human motions.
{"title":"Hierarchical motion history images for recognizing human motion","authors":"James W. Davis","doi":"10.1109/EVENT.2001.938864","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938864","url":null,"abstract":"There has been increasing interest in computer analysis and recognition of human motion. Previously we presented an efficient real-time approach for representing human motion using a compact \"motion history image\" (MHI). Recognition was achieved by statistically matching moment-based features. To address previous problems related to global analysis and limited recognition, we present a hierarchical extension to the original MHI framework to compute dense (local) motion flow directly from the MHI. A hierarchical partitioning of motions by speed in an MHI pyramid enables efficient calculation of image motions using fixed-size gradient operators. To characterize the resulting motion field, a polar histogram of motion orientations is described. The hierarchical MHI approach remains a computationally inexpensive method for analysis of human motions.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130009457","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938869
M. Petkovic, W. Jonker
As amounts of publicly available video data grow the need to query this data efficiently becomes significant. Consequently content-based retrieval of video data turns out to be a challenging and important problem. We address the specific aspect of inferring semantics automatically from raw video data. In particular, we introduce a new video data model that supports the integrated use of two different approaches for mapping low-level features to high-level concepts. Firstly, the model is extended with a rule-based approach that supports spatio-temporal formalization of high-level concepts, and then with a stochastic approach. Furthermore, results on real tennis video data are presented, demonstrating the validity of both approaches, as well us advantages of their integrated use.
{"title":"Content-based video retrieval by integrating spatio-temporal and stochastic recognition of events","authors":"M. Petkovic, W. Jonker","doi":"10.1109/EVENT.2001.938869","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938869","url":null,"abstract":"As amounts of publicly available video data grow the need to query this data efficiently becomes significant. Consequently content-based retrieval of video data turns out to be a challenging and important problem. We address the specific aspect of inferring semantics automatically from raw video data. In particular, we introduce a new video data model that supports the integrated use of two different approaches for mapping low-level features to high-level concepts. Firstly, the model is extended with a rule-based approach that supports spatio-temporal formalization of high-level concepts, and then with a stochastic approach. Furthermore, results on real tennis video data are presented, demonstrating the validity of both approaches, as well us advantages of their integrated use.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129185863","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938867
C. Rao, M. Shah
Automatically understanding human actions from video sequences is a very challenging problem. This involves the extraction of relevant visual information from a video sequence, representation of that information in a suitable form, and interpretation of visual information for the purpose of recognition and learning. We first present a view-invariant representation of action consisting of dynamic instants and intervals, which is computed using spatiotemporal curvature of a trajectory. This representation is then used by our system to learn human actions without any training. The system automatically segments video into individual actions, and computes a view-invariant representation for each action. The system is able to incrementally, learn different actions starting with no model. It is able to discover different instances of the same action performed by different people, and in different viewpoints. In order to validate our approach, we present results on video clips in which roughly 50 actions were performed by five different people in different viewpoints. Our system performed impressively by correctly interpreting most actions.
{"title":"View-invariant representation and learning of human action","authors":"C. Rao, M. Shah","doi":"10.1109/EVENT.2001.938867","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938867","url":null,"abstract":"Automatically understanding human actions from video sequences is a very challenging problem. This involves the extraction of relevant visual information from a video sequence, representation of that information in a suitable form, and interpretation of visual information for the purpose of recognition and learning. We first present a view-invariant representation of action consisting of dynamic instants and intervals, which is computed using spatiotemporal curvature of a trajectory. This representation is then used by our system to learn human actions without any training. The system automatically segments video into individual actions, and computes a view-invariant representation for each action. The system is able to incrementally, learn different actions starting with no model. It is able to discover different instances of the same action performed by different people, and in different viewpoints. In order to validate our approach, we present results on video clips in which roughly 50 actions were performed by five different people in different viewpoints. Our system performed impressively by correctly interpreting most actions.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122603456","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 : 2001-07-08DOI: 10.1109/EVENT.2001.938860
M. Harville, G. Gordon, J. Woodfill
Segmentation of novel or dynamic objects in a scene, often referred to as "background subtraction" or foreground segmentation", is a critical early in step in most computer vision applications in domains such as surveillance and human-computer interaction. All previously described, real-time methods fail to handle properly one or more common phenomena, such as global illumination changes, shadows, inter-reflections, similarity of foreground color to background and non-static backgrounds (e.g. active video displays or trees waving in the wind). The advent of hardware and software for real-time computation of depth imagery makes better approaches possible. We propose a method for modeling the background that uses per-pixel, time-adaptive, Gaussian mixtures in the combined input space of depth and luminance-invariant color. This combination in itself is novel, but we further improve it by introducing the ideas of (1) modulating the background model learning rate based on scene activity, and (2) making color-based segmentation criteria dependent on depth observations. Our experiments show that the method possesses much greater robustness to problematic phenomena than the prior state-of-the-art, without sacrificing real-time performance, making it well-suited for a wide range of practical applications in video event detection and recognition.
{"title":"Foreground segmentation using adaptive mixture models in color and depth","authors":"M. Harville, G. Gordon, J. Woodfill","doi":"10.1109/EVENT.2001.938860","DOIUrl":"https://doi.org/10.1109/EVENT.2001.938860","url":null,"abstract":"Segmentation of novel or dynamic objects in a scene, often referred to as \"background subtraction\" or foreground segmentation\", is a critical early in step in most computer vision applications in domains such as surveillance and human-computer interaction. All previously described, real-time methods fail to handle properly one or more common phenomena, such as global illumination changes, shadows, inter-reflections, similarity of foreground color to background and non-static backgrounds (e.g. active video displays or trees waving in the wind). The advent of hardware and software for real-time computation of depth imagery makes better approaches possible. We propose a method for modeling the background that uses per-pixel, time-adaptive, Gaussian mixtures in the combined input space of depth and luminance-invariant color. This combination in itself is novel, but we further improve it by introducing the ideas of (1) modulating the background model learning rate based on scene activity, and (2) making color-based segmentation criteria dependent on depth observations. Our experiments show that the method possesses much greater robustness to problematic phenomena than the prior state-of-the-art, without sacrificing real-time performance, making it well-suited for a wide range of practical applications in video event detection and recognition.","PeriodicalId":375539,"journal":{"name":"Proceedings IEEE Workshop on Detection and Recognition of Events in Video","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2001-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131836431","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：