123 results about "Multitree" patented technology

Coding schemes for coding a spatially sampled information signal using sub-division and coding schemes for coding a sub-division or a multitree structure are described, wherein representative embodiments relate to picture and / or video coding applications.

A better compromise between encoding complexity and achievable rate distortion ratio, and / or to achieve a better rate distortion ratio is achieved by using multitree sub-divisioning not only in order to subdivide a continuous area, namely the sample array, into leaf regions, but using the intermediate regions also to share coding parameters among the corresponding collocated leaf blocks. By this measure, coding procedures performed in tiles—leaf regions—locally, may be associated with coding parameters individually without having to, however, explicitly transmit the whole coding parameters for each leaf region separately. Rather, similarities may effectively exploited by using the multitree subdivision.

A favorable merging or grouping of simply connected regions into which the array of information samples is sub-divided, is coded with a reduced amount of data. To this end, a predetermined relative locational relationship is defined enabling an identifying, for a predetermined simply connected region, of simply connected regions within the plurality of simply connected regions which have the predetermined relative locational relationship to the predetermined simply connected region. Namely, if the number is zero, a merge indicator for the predetermined simply connected region may be absent within the data stream. In other embodiments, spatial sub-division is performed depending on a first subset of syntax elements, followed by combining spatially neighboring simply connected regions depending on a second subset of syntax elements, to obtain an intermediate sub-division.

The invention belongs to the field of artificial intelligence and particularly relates to a hop searching based conflict-avoiding path planning method for multiple robots. The method specifically comprises the following steps: 1. preprocessing a map into a data map; 2. carrying out path planning on all agents by using a searching framework of CBS, i.e., Conflict-based Search and using an unconstrained hop searching algorithm; 3. traversing a multiway tree to carry out conflict detection, constraint increasing and constrained path planning on all the agents. According to the hop searching basedconflict-avoiding path planning searching algorithm, provided by the invention, through greatly reducing the quantity of expanding nodes during path planning, the efficiency of searching of conflict-free path sequence algorithms of all the agents planned by a system is increased; during path re-searching in conflict processing, real distance from a current node to an end point is directly searched through an EfftiveG table, thus, the time for planning conflict-free path sequences of all the agents by the system is shortened again, and the searched shortest path can be closer to a real path.

The invention discloses an indexing method applied to time sequence historical database, which is used for realizing the high-efficiency storage and retrieval of the time sequence historical database. According to the indexing method, the time point is taken as a keyword to establish multi-tree indexes, and the indexes can be orderly stored in an index file in a disc according to the time sequence, and the time sequence historical database can be orderly stored in the data file in a disc according to the time sequence. During data access, the access frequency of files can be reduced through batching memory mapping file mechanisms, so that the data storage and retrieval efficiency of the time sequence historical database can be ensured, and moreover, the consumption on memory resources is small.

The invention provides a route prefix processing, lookup, adding and deleting method. A multi-way Trie tree is utilized to artfully design a Trie node data structure, and the technical problems that the traditional routing table realization method can not satisfy the quick route lookup requirement, the memory access times are reduced, the cost of construction is high and the use efficiency is low can be solved by a prefix list item storage data structure and algorithm, a high-speed longest route prefix lookup method, a quick route prefix increasing method and a quick route prefix deleting method. The route prefix processing, lookup, adding and deleting method has the advantages of high lookup speed, high update speed and good use benefit.

The invention provides a hybrid spatial indexing mechanism for processing geographic text Skyline query, wherein geographic text Skyline query refers to performing Skyline query on a geographic text information dataset. Data points in the geographic text dataset include geographic position information and keyword text information. The Inverted-Merged R*-Tree (IMR*-T) integrates the R*tree and theInverted File thought, and the invention belongs to the query indexing field in computer science. The invention focuses on solution of the problem of storing the geographic text dataset and performingSkyline query on the geographic text dataset, can improve the Skyline query efficiency on the premise of ensuring reasonable storage. The invention constructs a multi-branch tree according to data point spatial position distribution by means of an R* tree construction strategy, and can construct an Inverted File for leaf nodes of the tree. To improve the clipping efficiency of the dataset, tree nodes store boundary frame information. The invention is widely suitable for relevant application scenarios of geographic text Skyline query.

The invention discloses a longitudinal federation learning calculation method and device, equipment and a medium, and relates to the field of federation learning in artificial intelligence. Accordingto the application, a plurality of participation nodes are deployed by adopting a multi-way tree topology, one upper-layer participation node is provided with k lower-layer participation nodes, and the upper-layer participation node and the k lower-layer participation nodes exchange public keys of two parties, the upper-layer participation node and the lower-layer participation node perform two-party joint security calculation by taking the first public key and the second public key as encryption parameters to obtain k two-party joint outputs of the federation model; and the k two-party jointoutputs are combined by the upper-layer participation node to obtain a first joint model output corresponding to the federation model. Therefore, a longitudinal federated learning architecture of multi-way tree topology deployment is provided, and the equivalence of each participating node in the longitudinal federated learning process is improved.

The invention discloses a financial big data-oriented multi-way tree structure block chain integrated optimization storage method. The method comprises the following steps of 1, constructing a multi-way tree structure block chain integrated optimization model; 2, constructing a private chain block based on a Merkle tree; 3, constructing an alliance chain block based on a Merkle Pattercia tree; and4, establishing a communication consensus mechanism of the block chain. By means of the method, the multi-way tree structure block chain integrated optimization storage method oriented to financial big data is provided, and compared with an existing storage method, the storage efficiency, expenditure and other aspects are obviously improved; and during query, the query efficiency and accuracy arealso remarkably improved, and the application requirements in the current big data era are met.

The invention discloses an event-driven state-oriented self-adaptive sensor organizational method which comprises the steps that 1) an one-event multi-mode response model is set, and an event is divided into a plurality of levels; 2) when node acquisition data are in a sampling value section corresponding to an event of a certain level, an event message is structured to be broadcast to the internet of things; 3) after the event message is received by other nodes, if the local caches the event message is judged, if yes, the message is ignored, if not, the event message is cached in the local and the response level and report frequency of the event message are determined, and meanwhile the node identification and address are added to a message route list for subsequent broadcast; 4) a laminated multi-way tree structure is structured according to node levels, and a complete link is structured for the event; 5) the node in the complete link transmits the data of the event to the higher level node according to the data report frequency corresponding to the response level of the node until the data are sent to the sink node. According to the method, multi-mode response of nodes and network quick organization are achieved.

The invention discloses a block adaptive type Cartesian grid rapid graph mapping method and system, and the method comprises the following steps: S1, after an MPI environment is initialized, importing geometric data of a distributed block adaptive type Cartesian grid into each processor so as to reconstruct a multi-way tree structure, and building a link relation of adjacent hierarchical nodes; S2, sharing node geometric information at a region decomposition boundary across processors, and storing the node geometric information in a corresponding virtual node data container; S3, constructing a local geometric topology of each leaf node on the current processor; S4, normalizing the completely overlapped geometric elements on each processor; S5, based on cross-processor query operation, establishing dynamic communication mapping of docking surface elements at the region decomposition boundary; and S6, exporting the non-structural polyhedron computational grids in parallel on each processor. The problems that in the prior art, a self-adaptive geometric module and a physical field numerical simulator are difficult to fuse, and the dynamic geometric data format conversion efficiency is low are solved.

The invention provides a point cloud isolated point coding method and device and a point cloud isolated point decoding method and device. The method comprises the following steps of acquiring geometric information of ta o-be-coded point cloud; carrying out coordinate translation and coordinate quantification on the geometric position of the to-be-coded point cloud, so that the to-be-coded point cloud is contained in a bounding box; performing multi-way tree division on bounding box iteration based on the geometric position of the to-be-coded point cloud after coordinate quantification to obtain a current to-be-coded block of the layer after each division; for each current to-be-coded block, judging whether the current to-be-coded block meets a cut-off division condition or not based on thegeometric position of the point cloud point and the number of points; if the cut-off division condition is not met, continuing to execute the multi-way tree division; and encoding the point cloud point entropy in the current to-be-encoded block meeting the cut-off division condition into a binary code stream. According to the cut-off division condition provided by the invention, for the current to-be-coded block with repeated points, if the cut-off division condition is met, entropy coding is directly carried out without continuous division, so that the coding and decoding complexity is reduced.

According to the invention, an efficient message receiving and dispatching middleware model is designed and realized, a message rapid receiving and dispatching and multi-priority distribution engine fusing PUSH and PULL modes, a message queue persistence optimization strategy and a timing backup and fault-tolerant processing mechanism are included, and the problem of low judgment accuracy of message middleware under the condition of high network delay is solved. A node failure judgment algorithm based on a subjective conclusion is provided, a multi-way tree is improved to further optimize a consistent Hash algorithm, and the problem that node loads are not uniform in a traditional method is solved; and in order to ensure the consistency of data access in the cluster, an original distributed lock based on a single node is improved, and the safety of the distributed lock in the cluster is improved. Experimental results prove that the distributed network message queue middleware has higher judgment accuracy, more balanced cluster node load and better performance in message receiving and sending speed, and is more efficient, lightweight and easy to use.

The invention provides a different structure converting method from an AOE (Activity On Edge) network to a multi-way tree structure, and aims to provide a multidimensional MBOM (Manufacturing Bill of Material) converting method based on the AOE network, which is simple and visible and occupies fewer system resources. The different structure converting method is realized through the technical scheme including the following steps: constructing an MBOM; conducting topological sorting on each node in the AOE network chart adopting the topological sorting algorithm; reading all nodes adopting the FIFO (First-In First-Out) algorithm by a data register; recording the nodes as PP (Process Plan) and performing one-way acyclic graph topological sorting; and then, performing recursion and traversing on all paths of each node through an MBOM conversion adapter adopting the DFS (Depth-First Search) algorithm; calculating the sum of the weights of all sides of each path; obtaining the path with the maximum weight sum and storing the path in the database; taking the path with the maximum weight sum as the critical path and the main path; inverting sequential relationship of all paths to generate an MBOM multi-way tree structure; reading the nodes layer-by-layer from top to bottom; writing the father node into a relational database table.

The invention discloses a time sequence-based multiway tree data index structure cloud storage privacy protection method. The method comprises the following process that by creating a time sequence-based multiway tree data storage structure (MTTS) on a cloud trusted center, effective management is conducted on data names, and the retrieval speed is increased. A key derivation algorithm is designedon the basis of the MTTS structure to achieve key management and distribution, an encrypted keyword retrieval algorithm is constructed to support data sharing and ciphertext retrieval, a data privacyprotection method is designed on a proposed cloud storage architecture for the two processes of data storage and data reading, and the user access authority changing and data updating problems are solved by combining a delayed updating technology. According to the design of the cloud storage privacy protection method, the privacy of a user can be effectively protected, and efficient data access can be supported.

The invention provides a rendering method and device in combination with audios and videos, a computer readable storage medium and computer equipment, and the method comprises the steps: loading a scene, processing the scene through employing a multi-way tree, and obtaining a multi-way tree data structure comprising a plurality of first rendering units; reading current video frame data obtained after the audio and video file is decoded, forming a second rendering unit by the current video frame data, and adding the second rendering unit into the multi-way tree data structure; traversing the multi-way tree data structure, and screening out an effective rendering unit from the first rendering unit and the second rendering unit according to a preset rule; and submitting the effective rendering unit to a GPU for rendering. According to the method, the combination of audio and video stream playing and graphic rendering is completed on the basis of realizing high-performance graphic rendering, the video rendering effect is realized, the requirement of high-definition video playing is met, and the expressive force of a video picture is enhanced.

The invention discloses a multi-box-type three-dimensional boxing optimization method and system based on multi-way tree search. The method comprises the following steps: carrying out model building, wherein a multi-box-type three-dimensional boxing optimization model is built; constructing a multi-way tree search structure: constructing the multi-way tree search structure for solving a multi-box type three-dimensional boxing optimization problem; optimization algorithm: in the multi-way tree search process, calling a single-box three-dimensional boxing optimization algorithm once when nodes are expanded and generated each time, and obtaining an optimal single-box loading scheme; in the multi-way tree searching process, adopting a heuristic pruning strategy and a funnel type node number control strategy optimization algorithm; and outputting a result. The invention aims to provide a multi-box type three-dimensional boxing optimization method and system based on multi-way tree search, a current optimal boxing scheme is obtained by calling a single-box three-dimensional boxing optimization algorithm, the search scale of a tree is reduced by applying a heuristic pruning strategy and a funnel type node number control strategy, the search efficiency is improved, a large amount of redundant search is reduced, and the calculation time is shortened.

The invention discloses a tree model-based complex natural language query-to-SQL method, which comprises the following steps: firstly, pre-screening an input database to obtain TOP K tables related to the natural language query, converting multi-table query into single-table query, so as to suit for the situation that information cannot be input into the natural language query-to-SQL model due to the fact that the database scale is large; secondly, disassembling the complex SQL statement into a multi-way tree form from top to bottom, each node representing a module, and therefore a tree-shaped model is constructed to predict output of each module of the complex SQL statement; and finally, aiming at an SQL statement value extraction task, separating value extraction from the tree model, and sharing weights to extract values of different clauses in the SQL at the same time. According to the tree model-based complex natural language query-to-SQL method framework, the problem of multi-table complex query in a natural language query-to-SQL task can be solved.

A better compromise between encoding complexity and achievable rate distortion ratio, and / or to achieve a better rate distortion ratio is achieved by using multitree sub-divisioning not only in order to subdivide a continuous area, namely the sample array, into leaf regions, but using the intermediate regions also to share coding parameters among the corresponding collocated leaf blocks. By this measure, coding procedures performed in tiles—leaf regions—locally, may be associated with coding parameters individually without having to, however, explicitly transmit the whole coding parameters for each leaf region separately. Rather, similarities may effectively exploited by using the multitree subdivision.

A better compromise between encoding complexity and achievable rate distortion ratio, and / or to achieve a better rate distortion ratio is achieved by using multitree sub-divisioning not only in order to subdivide a continuous area, namely the sample array, into leaf regions, but using the intermediate regions also to share coding parameters among the corresponding collocated leaf blocks. By this measure, coding procedures performed in tiles—leaf regions—locally, may be associated with coding parameters individually without having to, however, explicitly transmit the whole coding parameters for each leaf region separately. Rather, similarities may effectively exploited by using the multitree subdivision.

A better compromise between encoding complexity and achievable rate distortion ratio, and / or to achieve a better rate distortion ratio is achieved by using multitree sub-divisioning not only in order to subdivide a continuous area, namely the sample array, into leaf regions, but using the intermediate regions also to share coding parameters among the corresponding collocated leaf blocks. By this measure, coding procedures performed in tiles—leaf regions—locally, may be associated with coding parameters individually without having to, however, explicitly transmit the whole coding parameters for each leaf region separately. Rather, similarities may effectively exploited by using the multitree subdivision.

The invention discloses a data analysis method which comprises the steps of constructing a recommend data set based on an ordered multi-way tree, wherein each of nodes of the ordered multi-way tree represents a recommend datum and weight of branches for enabling all the nodes to be connected represents correlation among different recommend data; receiving first selection data from a user; determining the position of the first selection data in the ordered multi-way tree; with the position where the first selection data is positioned serving as a father node, performing depth or breadth traversal in the recommend data set based on the ordered multi-way tree in order to output one or more appropriate recommend data to the user. The invention further discloses a data analysis device.

The invention discloses a three-dimensional tree model real-time simplification method based on viewpoint mutual information. The method can enable rendering of trees in a forest scene to be smootherand obtain a better visualization effect, and specifically comprises the following steps: dividing the trees into nodes with father-son relationships according to topological relationships among branches, and organizing the nodes in a form of a multi-way tree; calculating the visual importance of each leaf in each leaf node according to the viewpoint mutual information VMI, and reordering the leaves according to the visual importance; and proposing a plurality of factors influencing the real-time simplification rate of the nodes, and setting a corresponding function for each factor to determine the real-time simplification rate of the nodes. Optimization measures for rendering a large-scale forest scene, including LOD switching control, distant view crown replacement rendering and the like, are provided. By means of VMI-based leaf sorting, the visual appearance characteristics of trees can be well kept while the high simplification rate is achieved. In addition, better visualization effect and rendering efficiency are ensured by optimization measures in the rendering process.

The invention provides a transformer area topological relation graph automatic drawing method based on multi-way tree layered layout, being characterized in that automatic layout is carried out by importing a topological file, so that manual drawing is not needed, and the working efficiency is high; coordinates of a main graph are calculated layer by layer, and the calculation complexity is reduced; an automatic mapping result is divided into a main graph and a plurality of sub-graphs, and the main graph and the sub-graphs can be carried out in parallel, and the mapping speed is high, and themain topology of the current transformer area can be presented more visually; when the automatic mapping result is not satisfactory, the manual adjustment can be realized, and the flexibility is higher; an automatic mapping result has no intersection of connecting lines and is clearer; and sub-graph electric energy meters are drawn in a split-phase mode, and inspectors can be helped to finely manage household meters.

The invention discloses a sensor network fault active sensing method and system. The method comprises the steps that a multi-way tree of a topological structure in a cluster area is constructed according to a topological graph generated by clustered nodes; each leaf node sequentially judges whether the currently communicating leaf node is a fault node or not according to a consensus mechanism; afault node is marked, and a sub-tree which does not contain the fault node is regenerated; and a multi-way tree is formed in the clustering area through the consensus mechanism, so that the stabilityof the quality of the data acquired by each partition in the wireless sensor network is improved. The fault area is determined to be limited in the sub-tree of each multi-way tree, so that the energyconsumption is greatly reduced. Compared with other full-area fault node elimination methods through fault sensing, the method has the advantages that the network survival time is prolonged, the areaswhere the fault nodes appear can be roughly known through a base station, and therefore the areas where the fault nodes exist can be rapidly positioned.

The invention discloses a text classification method and device based on a heterogeneous neural network, and the method comprises the steps: S1, constructing an N-layer text classification multi-way tree corresponding to a tree structure in combination with the tree structure of an actual classification system in a training data set, and according to the structure of the N-layer text classification multi-way tree, respectively writing training data in the training data set into classification files corresponding to each level, performing word segmentation on Chinese texts of each classification file, performing feature selection, and storing the selected features into corresponding feature files; S2, constructing a text heterogeneous neural network corresponding to the tree structure; and S3, setting heterogeneous neural network algorithm parameters, carrying out parameter adjustment, iteratively generating and storing each classification and sub-classification model of training data, using a verification data set for carrying out accuracy judgment. The categories and the hierarchical relation between the categories are also added into the heterogeneous neural network, and the problems of vector representation and link data sparsity are learned in a category display manner.

The embodiment of the invention discloses a multi-way tree forest K-anonymity method and system, equipment and a readable storage medium. On the basis of calculation of the recognition rate, the problem of each step of optimal generalization data selection of a greedy algorithm in a local generalization mode is solved, a multi-way tree forest is established to achieve storage and local generalization of generalization data, and the efficiency of the model is improved through multi-thread generalization data selection and K-anonymity parallel calculation judgment. That is to say, part of data is generalized on the basis of original data, so that an anonymous data set meets a K-anonymity model. The authenticity of the data is not influenced under the condition of ensuring the data security.Meanwhile, the model reserves the distribution of the data, improves the availability of the data after data anonymization, and is suitable for anonymization processing of a small data set and a largedata set.

The invention discloses a layered iterative layout optimization method considering cutting efficiency and utilization rate. The method comprises the following steps: establishing a model: establishinga multi-objective optimization model for minimizing the total cost in combination with the cutting efficiency and the utilization rate; slice typesetting is conducted, specifically, stacked slices are completed through a layout recursive algorithm based on stacked plate number layering; setting a threshold value: setting the threshold value based on a multi-way tree hierarchical iterative searchmethod of a utilization rate threshold value; analyzing and optimizing parameters: analyzing and optimizing the parameters obtained by hierarchical iterative search; calculation experiment and analysis: testing and analyzing through experiments to verify the feasibility and effectiveness of algorithm results; the invention aims to provide the layered iterative layout optimization method considering the cutting efficiency and the utilization rate, the utilization rate of raw materials is improved, batch cutting can be realized, and the cutting frequency and the cutting knife number can be greatly reduced, so that the raw material cost is reduced, the production efficiency is improved, and the total production cost is reduced.