detection of anomalies in multivariate time series is an important data mining task with potential applications in medical diagnosis, ecosystem modeling, and network traffic monitoring. In this paper, we present a rob...
详细信息
detection of anomalies in multivariate time series is an important data mining task with potential applications in medical diagnosis, ecosystem modeling, and network traffic monitoring. In this paper, we present a robust graph-based algorithm for detecting anomalies in noisy multivariate time series data. A key feature of the algorithm is the alignment of kernel matrices constructed from the time series. The aligned kernel enables the algorithm to capture the dependence relationship between different time series and to support the discovery of different types of anomalies (including subsequence-based and local anomalies). We have performed extensive experiments to demonstrate the effectiveness of the proposed algorithm. We also present a case study that shows the utility of applying our algorithm to detect ecosystem disturbances in Earth science data.
Deadlock detection is a well-studied problem that may be considered solved from a theoretical point of view. However, specific cases may demand for specific solutions. One such specific case is deadlock detection in K...
详细信息
Deadlock detection is a well-studied problem that may be considered solved from a theoretical point of view. However, specific cases may demand for specific solutions. One such specific case is deadlock detection in Kahn Process Networks. The Kahn process network (KPN) is an expressive model of computation that is widely used to model and specify deterministic streaming applications. The processes in the network communicate point-to-point over FIFO channels whose sizes are undecidable in general. As a consequence, deadlock may occur and, therefore, a run-time deadlock detection mechanism is required. This can be organized in a centralized way, a distributed way, and a hierarchical way. Centralized and distributed procedures have been reported in the literature. In this paper, we propose a novel hierarchical approach for KPN deadlock detection at run time. We also give results for the implementation on the IBM Cell processor.
The storage operation of normal process in host system is analyzed and an anomaly intrusion detection method based on d-s evidence theory for storage system is proposed. The detector fuses multiple signatures of stora...
详细信息
The storage operation of normal process in host system is analyzed and an anomaly intrusion detection method based on d-s evidence theory for storage system is proposed. The detector fuses multiple signatures of storage data to decide whether the storage operation flow is normal. Furthermore, six groups of light-computation signatures of storage operation data are used to develop an efficient fusion mechanism to guarantee high performance of the algorithm. Experiment shows that high detection rate can be achieved by such fusion.
Dissolves are the basic artificial effects of gradual transitions in video sequences, which are widely used in TV programs and movies. Through dissolve detection, interesting and important video temporal segments can ...
详细信息
Dissolves are the basic artificial effects of gradual transitions in video sequences, which are widely used in TV programs and movies. Through dissolve detection, interesting and important video temporal segments can be easily located and indexed for various applications. In this paper, we present an effective dissolve detection algorithm, which takes into account the dissolve properties in both temporal and spatial domains. In particular, in the temporal domain, we use frame difference to capture the dissolve characteristics. In the spatial domain, the central area is given bigger weight than the four sides. Experimental results show a good performance of our proposed algorithm.
Concerning the requirements of real-time and accurate collision detection in interactive system, we propose a shared memory parallel collision detection algorithm. First we incorporate the merits of both AABB bounding...
详细信息
Concerning the requirements of real-time and accurate collision detection in interactive system, we propose a shared memory parallel collision detection algorithm. First we incorporate the merits of both AABB bounding box and bounding spheres to construct a hybrid bounding representation of arbitrary non-convex polyhedra (S-AABB) for attaining speed, and then use OpenMP parallel programming model to traversal the built hybrid bounding volume hierarchy, so further accelerate the collision detection algorithm. At last, experiments results have shown that our algorithm is advantageous over other current typical collision detection algorithms such as I-COLLIDE [1] regarding efficiency and accuracy, so can meet the real-time and accurate requirements in complex interactive virtual environment.
In multiuser detection, the set of users active at any time may be unknown to the receiver. A two-step detection procedure, in which multiuser detection is preceded by active-user identification, is suboptimum. The op...
详细信息
In multiuser detection, the set of users active at any time may be unknown to the receiver. A two-step detection procedure, in which multiuser detection is preceded by active-user identification, is suboptimum. The optimum solution consists of detecting simultaneously the set of active users and their data, problem that can be solved exactly by applying random-set theory (RST). However, implementation of optimum detectors can be limited by their complexity, which grows exponentially with the number of potential users. In this paper we illustrate how the complexity of optimum can be reduced. In particular, sphere detection (SD) techniques (possibly in an approximate version) are examined.
Citation graph analysis has been used to evaluate the significance of documents and authors, or to estimate the impact of publication venues. In this paper, we investigate its new application in topic identification. ...
详细信息
Citation graph analysis has been used to evaluate the significance of documents and authors, or to estimate the impact of publication venues. In this paper, we investigate its new application in topic identification. We first model the communities in the citation graph as related documents on a specific topic. And then, a scientific theme detection algorithm is proposed based on community partition, attempting to identify the emergency of a new theme by tracking the change of the community where the top cited nodes lie in. Experimental results on real dataset show that the proposed method can detect new topic timely with only a subset of data.
In this paper, a detection algorithm with parallel partial candidate-search algorithm is presented. Two fully independent partial search processes are simultaneously employed for two groups of transmit antennas based ...
详细信息
In this paper, a detection algorithm with parallel partial candidate-search algorithm is presented. Two fully independent partial search processes are simultaneously employed for two groups of transmit antennas based on QR and QL decompositions of the channel matrix. Proposed QRD- QLD detection algorithm is compared with well-known QRD-M scheme adopted for several emerging wireless standards. Latency of the QRD-QLD candidate search is about twice as small for similar error-rate performance and for identical hardware resources. Total detection latency of QRD-QLD algorithm that also includes computation of soft information for outer decoder is also substantially smaller.
In this paper, a road detection method based on an image segmentation algorithm is presented. Road detection process is a key issue for an autonomous driving system in urban environment. Image-based road detection alg...
详细信息
In this paper, a road detection method based on an image segmentation algorithm is presented. Road detection process is a key issue for an autonomous driving system in urban environment. Image-based road detection algorithm is applied on sources of visual information recorded by cameras when our car is running on road. Our method combines a posteriori probability and visual information for image segmentation. The method is composed of two steps. Firstly, a road identifier is trained with supervised learning algorithm. Secondly, road regions are detected by combining a posteriori probability and visual information using image segmentation algorithm. Experimental results are presented for road images in actual driving video acquired in urban areas.
Increasing trend in the usage of translucent television logos by broadcast channels renders opaque logo detection algorithms inadequate. Important applications such as identification of broadcast types make the detect...
详细信息
Increasing trend in the usage of translucent television logos by broadcast channels renders opaque logo detection algorithms inadequate. Important applications such as identification of broadcast types make the detection of translucent logos a significant requirement. This paper presents a method for detecting translucent television logos in video streams. Firstly, boundary information of the logo, which will be searched in broadcast stream, is extracted manually. This search is carried out by comparing the edge map of the luminance channel of the interest region with inner and outer contours of the logo using different metrics. Performance is increased by utilization of temporal redundancies and solutions to special problematic cases. Furthermore, traces of the logo boundaries are examined in chrominance channels of video frames in order to eliminate false alarms caused by opaque logos with the same boundaries. Promising results indicate the applicability of the method in real life.
暂无评论