Human motion recognition plays a crucial role in the video analysis framework.However,a given video may contain a variety of noises,such as an unstable background and redundant actions,that are completely different fr...Human motion recognition plays a crucial role in the video analysis framework.However,a given video may contain a variety of noises,such as an unstable background and redundant actions,that are completely different from the key actions.These noises pose a great challenge to human motion recognition.To solve this problem,we propose a new method based on the 3-Dimensional(3D)Bag of Visual Words(BoVW)framework.Our method includes two parts:The first part is the video action feature extractor,which can identify key actions by analyzing action features.In the video action encoder,by analyzing the action characteristics of a given video,we use the deep 3D CNN pre-trained model to obtain expressive coding information.A classifier with subnetwork nodes is used for the final classification.The extensive experiments demonstrate that our method leads to an impressive effect on complex video analysis.Our approach achieves state-of-the-art performance on the datasets of UCF101(85.3%)and HMDB51(54.5%).展开更多
As the means for communication of users in different communication networks and network resource sharing, communication network interconnection has been paid great attention to in recent years. The performance analysi...As the means for communication of users in different communication networks and network resource sharing, communication network interconnection has been paid great attention to in recent years. The performance analysis of internetworks is very important in communication network interconnection, for it is the basis for the study of other problems concerned. The main contribution of this paper is a wide and deep study of performance of internetwork, the construction of systematic performance analysis method for arbitrary internetworks, and of queuing model for gateways. In this paper, the performance analysis of internetwork with arbitrary topology is studied for the first time, and a new method - Node Equivalence of Subnetworks (NES) is proposed. The key step of this method is the node equivalence of subnetworks of all kinds, so we first deeply study the node equivalence, and develop two models for node equivalence. Then we study the node equivalence of CSMA/CD and SLOTTED ALOHA systems by using these models. The results show that the NES method is a systematic method which is suitable for performance analysis of any type of internetworks. The proposed equivalence models is usable with good approximation. The queuing model of gateway is first widely studied in this paper. The queuing model for gateways with packet fragmentation and reassembly is studied for the first time and the analysis method, proposed. In the analysis, two modeling tools, Reversed Switched Poisson Process (RSPP) and Switched Determined Process (SDP), are proposed and RSPP/M/1 queue and SDP/D/1 queue, analyzed. A new method-minislot method is proposed and used for the study of SDP/D/1 queue and performance of ATM multiplexer. In this paper, the interdeparture time distribution of ATM network under some input traffic models is studied and the expression, derived. These results can be used for both performance evaluation of ATM network and the performance of ATM networks interconnected with other networks. We also propose an approximate analysis method for queuing networks with priority service and the queuing networks with packet fragmentation and reassembly.展开更多
文摘Human motion recognition plays a crucial role in the video analysis framework.However,a given video may contain a variety of noises,such as an unstable background and redundant actions,that are completely different from the key actions.These noises pose a great challenge to human motion recognition.To solve this problem,we propose a new method based on the 3-Dimensional(3D)Bag of Visual Words(BoVW)framework.Our method includes two parts:The first part is the video action feature extractor,which can identify key actions by analyzing action features.In the video action encoder,by analyzing the action characteristics of a given video,we use the deep 3D CNN pre-trained model to obtain expressive coding information.A classifier with subnetwork nodes is used for the final classification.The extensive experiments demonstrate that our method leads to an impressive effect on complex video analysis.Our approach achieves state-of-the-art performance on the datasets of UCF101(85.3%)and HMDB51(54.5%).
文摘As the means for communication of users in different communication networks and network resource sharing, communication network interconnection has been paid great attention to in recent years. The performance analysis of internetworks is very important in communication network interconnection, for it is the basis for the study of other problems concerned. The main contribution of this paper is a wide and deep study of performance of internetwork, the construction of systematic performance analysis method for arbitrary internetworks, and of queuing model for gateways. In this paper, the performance analysis of internetwork with arbitrary topology is studied for the first time, and a new method - Node Equivalence of Subnetworks (NES) is proposed. The key step of this method is the node equivalence of subnetworks of all kinds, so we first deeply study the node equivalence, and develop two models for node equivalence. Then we study the node equivalence of CSMA/CD and SLOTTED ALOHA systems by using these models. The results show that the NES method is a systematic method which is suitable for performance analysis of any type of internetworks. The proposed equivalence models is usable with good approximation. The queuing model of gateway is first widely studied in this paper. The queuing model for gateways with packet fragmentation and reassembly is studied for the first time and the analysis method, proposed. In the analysis, two modeling tools, Reversed Switched Poisson Process (RSPP) and Switched Determined Process (SDP), are proposed and RSPP/M/1 queue and SDP/D/1 queue, analyzed. A new method-minislot method is proposed and used for the study of SDP/D/1 queue and performance of ATM multiplexer. In this paper, the interdeparture time distribution of ATM network under some input traffic models is studied and the expression, derived. These results can be used for both performance evaluation of ATM network and the performance of ATM networks interconnected with other networks. We also propose an approximate analysis method for queuing networks with priority service and the queuing networks with packet fragmentation and reassembly.
