47 results about "Disjoint-set" patented technology

Systems and methods are described that facilitate generating a background image with a UV-fluorescent watermark for printing on a document. A binary watermark mask is generated to separate the background image into the UV-active and the UV-dull regions. Based on the assigned binary value of the watermark mask, each pixel is assigned a UV-active or UV-dull color using an adaptive halftoning technique, in order to generate a binary UV-active image and a binary UV-dull image. A binary watermarked background image is generated by combining the binary UV-active and UV-dull images, and is printed on a document. The UV-active and UV-dull colors have different UV intensities under UV light, but are indistinguishable under normal lighting conditions. In this manner, the background image is visible when exposed to visible light, and the UV-fluorescent watermark is visible when exposed to UV light.

A computer implemented method of reducing central processing unit (CPU) usage of a firewall by safe reordering a current firewall's rule-base exhibiting N rules. The method comprising: receiving rule usage statistics exhibiting usage frequency of each rule on the current firewall's rule-base; calculating a rules matched per packet (RMPP) parameter, being a summation of products of each rule identifier and the corresponding usage frequency for all the N rules; determining an alternative order of the rule base by repositioning rules, wherein the repositioned rules perform the same action on the firewall, or wherein the repositioned rules act on disjoint sets of network connections, and wherein the repositioning results in a reduction of the RMPP of the reordered rule base, thereby reducing the CPU usage of the firewall in implementing the alternative order of rules.

A frequency partitioning technique is introduced that amortizes the work of computing codeword lengths within a tuplecode by grouping together tuples that have the same pattern of codeword lengths. Specifically, the technique entropy codes and partitions column values in each column into disjoint sets called column partitions, assigns a codeword length to each of the column partitions, identifies cells (a combination of codeword lengths), and collectively storing tuples associated with each of the cells.

The invention relates to a traffic playback method and system for a virtual network. The traffic playback method includes a first step of capturing and processing real traffic, extracting a real IP address set R_IP, a second step of conducting depth-first search on a bipartite graph which is generated by real traffic communication relationships, dividing the real IP address set R_IP into two disjoint sets, namely, a set R_IPA and a set R_IPB, a third step of dividing all virtual nodes which are in communication through any virtual network routing interface v_interfacei into two disjoint sets, namely a set V_IPAi and a set V_IPBi, a fourth step of calculating similarities of all the virtual network routing interfaces and a real traffic collecting point, a fifth step of selecting a virtual network interface which is most similar to the real traffic collecting point to be used as a mapping node of the traffic collecting point, conducting IP address mapping based on the mapping mode, and a sixth step of traversing the real traffic again to achieve real traffic playback in the virtual network. When the traffic is played back in the virtual network through the traffic playback method and system for the virtual network, the real traffic communication environment is restored as good as possible, and the virtual network traffic system is improved.

A vectoring controller for configuring a vectoring processor that jointly processes DMT communication signals to be transmitted over, or received from, a plurality of N subscriber lines according to a vectoring matrix. In accordance with an embodiment, the vectoring controller is configured, for given ones of a plurality of tones, to enable the given tone for direct data communication over a first set of N−Mk targeted lines out of the plurality of N subscriber lines, and to disable the given tone for direct data communication over a second disjoint set of Mk supporting lines out of the plurality of N subscriber lines, Mk denoting a non-null positive integer. The vectoring controller is further configured to configure the vectoring matrix to use an available transmit or receive power at the given tone over the second set of Mk supporting lines for further enhancement of data signal gains at the given tone over the first set of N−Mk targeted lines.

Coherent motion regions extend in time as well as space, enforcing consistency in detected objects over long time periods and making the algorithm robust to noisy or short point tracks. As a result of enforcing the constraint that selected coherent motion regions contain disjoint sets of tracks defined in a three-dimensional space including a time dimension. An algorithm operates directly on raw, unconditioned low-level feature point tracks, and minimizes a global measure of the coherent motion regions. At least one discrete moving object is identified in a time series of video images based on the trajectory similarity factors, which is a measure of a maximum distance between a pair of feature point tracks.

The invention provides a traffic sensing routing tree construction method based on geographic location routing. The method comprises the steps that a disjoint set data structure is used to track the membership of each node in a cluster; a priority queue data structure is used to store virtual bandwidths and distances to all adjacent clusters; the virtual bandwidth of two adjacent clusters is the sum of all available bandwidths between two cluster nodes, and a pair of adjacent clusters with the maximum attraction are merged first; the cluster head of a cluster with many nodes appears as a new cluster head; the cluster center is the average of the locations of all nodes in the cluster; the virtual bandwidth of the new cluster to the adjacent cluster is calculated from the bandwidth of the original cluster to the adjacent cluster; the highest available bandwidth link from the new cluster to the adjacent cluster is the highest bandwidth link from the original cluster to the adjacent cluster; and until only one cluster left or the rest of the clusters have no attraction, the method ends, and a large cluster with all nodes of a routing tree is acquired, wherein the cluster head of the large cluster is the node of the routing tree root. According to the invention, location and traffic information are considered at the same time; the number of collision packets and the average path hops are reduced; and the average path throughput is increased.

Lighting systems, devices, methods and apparatus for simulating a candle flame are disclosed. The systems and methods can employ spatially distributed, disjointed sets of light-emitting elements that are independently controlled (526) with randomized control signals to simulate a candle flame by forming spatially and temporally varying lighting effects. Further, the disjointed sets can be controlled (526) such that the total light intensity is essentially maintained. In addition, systems and methods can employ a transition probability model (508) to adjust intensities of light-emitting elements in a way that accurately mimics a candle flame.

A passive optical network (PON) has an optical line termination (OLT) that terminates an optical fiber servicing a plurality of optical network units (ONUs). Each ONU has one or more traffic containers (TCONTs) addressable by the OLT. The PON dynamic bandwidth allocation (DBA) implements a scheduling hierarchy, including several scheduling layers, such that disjoint sets of TCONTs can be grouped together, then disjoint sets of groups can be grouped, and so on. In such hierarchy, the residential traffic can be grouped separately from the business traffic. Further, within either the residential or business group, traffic may be grouped to define scheduling layers (“sub-groups”) within the residential or business group. Scheduling in one group or sub-group is performed independently of the scheduling in other groups or sub-groups, subject to the available bandwidth for each group. The scheduling may be controlled to allow the residential services to be oversubscribed while still ensuring compliance of service level performance metrics for the business services.

The invention discloses a two-stage behavior recognition fine classification method based on a graph convolutional network. According to the method, the accuracy of behavior recognition is improved mainly by reclassifying a difficult category set, and the method is divided into three stages: in the first stage, training a coarse classification model; in the second stage, acquiring difficult category sets and training difficult category set models, acquiring the difficult category sets by a confusion matrix of the rough classification model on the test set and a union-check set algorithm, and training the difficult category set models for the different difficult category sets; and in a third stage: during on-line inference, according to an inference result of the rough classification model, inputting the samples which need to be finely classified into the difficult category set model for re-classification. According to the method provided by the invention, aiming at the problem that a model is difficult to classify similar actions, the problem that the similar actions are difficult to classify is relieved to a certain extent by a rough classification-fine classification two-stage architecture, the accuracy of behavior recognition is improved, and a better result is obtained on a public data set.

In a ring having first and second adjacent nodes, and two fibers carrying information between the first and second nodes, the first fiber carries information in one direction, while the second fiber carries information in another direction. Each fiber includes wavelength capacity allocated to working and protection traffic. The working and protection wavelength capacities in the first fiber are respectively assigned to first and second disjoint sets of wavelengths, while the working and protection wavelength capacities in the second fiber are respectively assigned to the second and first disjoint sets of wavelengths.

A reduced complexity Multiprotocol Label Switching (MPLS) method, 30, a MPLS network element, 40, and a MPLS network, 10, utilize an MPLS operating regime whereby disjoint sets of one or more MPLS labels are uniquely and specifically associated with just one switch, i.e. each switch node is assigned one or more non-overlapping labels from the RFC 3032 20 bit label space to bind to particular service end-points; which then enables these labels to embody the core properties of a destination address (DA) in the network sub-domain in which they are used. The central property is that these DA labels are constant for a given forwarding path across the entire sub-domain, remaining unchanged at any extends in the network. Once that is achieved, any and all hop-by-hop signaling protocols are unnecessary, since there is no need for label swapping, and the label-switching-node binding information can be flooded by interior routing protocols only.

A method for tracking objects in a first memory during a plurality of time intervals is provided, wherein the method comprises: updating, by one or more update threads, copying, by the one or more test threads, the one or more non-copied shadow objects to a second memory, determining, by the one or more test threads whether the step of copying the one or more non-copied shadow objects was successful, and updating, by the one or more test threads, the status variables of corresponding objects that correspond to the one or more non-copied shadow objects that were copied successfully, wherein subsequent bits of the status variables correspond to subsequent time intervals and wherein the one or more update threads and the one or more test threads operate on disjoint sets of bits of the status variables.