442 results about "Relational graph" patented technology

A content-based video story browsing system is disclosed including setting and representing relations between objects and places of the story. A content-based video story browsing system includes the steps of selecting objects and places as factors organizing contents of a video story and setting a relational graph relating the objects to corresponding places; setting a video story section and a representative scene as an events of contents development corresponding to the relations of the objects and the places; and displaying the video story section and the representative scene and its contents as browsing information corresponding to the relation selected from the relational graph.

The invention provides a method and a device for tracking crowds and counting pedestrian flow. The method includes acquiring real-time color images and depth images of scenes; preprocessing the images; carrying out difference operation on backgrounds of the depth images to obtain moving foreground depth images; detecting pedestrian head regions in the foreground depth images and dividing the pedestrian head regions from the foreground depth images; judging the head regions and removing non-head regions; matching and tracking pedestrian heads according to joint matching probability functions; creating head state spatial switching relational graphs to count the pedestrian heads. The method and the device have the advantages that the position of each pedestrian can be tracked and recorded by the aid of the method and the device, the speeds and the directions of the pedestrians in walking procedures can be computed, the crowds can be tracked under various conditions, and the pedestrian flow can be counted under the various conditions; the method and the device are free of influence of illumination conditions and illumination change; false alarm due to false objects can be filtered out; the device is stable in performance and high in speed, efficiency and accuracy.

The present invention discloses a method of deducing the structural attributes of the online social network users. The method comprises the steps of coding a plurality of attributes of the users into the structural combined attribute category vectors, carrying out the weighted random walk in a user node relational graph G to obtain a user node sequence set, then utilizing a word-to-vector tool Word2Vec to generate the real-value vector representation of each user node, and constructing a full-connection neural network model to train, when the user attributes are deduced, inputting the user node vector representation of which the attribute needs to be deduced in the trained neural network model, calculating the probability of each combined attribute category vector, and taking the vector having the maximum probability as the combined attribute category of the user. According to the present invention, the attribute information of partial users and a friend relation (or a concern relation) between the users just need to be extracted, the additional user behavior feature data does not need to be obtained, at the same time, the provided method fully utilizes the internal relation between the attributes and enables the attribute deduction accuracy to be improved.

The invention discloses a code static analysis-based data race detecting method and a system of the detecting method. The method comprises the following steps of: reading software to be detected, statically analyzing a source program of the software to be detected, and generating an abstract syntax tree, a control flow graph and a global function call graph of the software to be detected; on thatbasis, computing alias information in each function, outlet alias information among functions, lock assembly information, an access link of an access escapable variable quantity in each function and a thread building relational graph; computing a plurality of initialized sentence pair sets of a plurality of access nodes in every two threads; and gradually eliminating the sets according to the alias information, the lock assembly information and a concurrency relation to obtain a sentence pair which can finally have the data race. The detecting method and the detecting system can effectively detect the data race of a multi-thread program compiled by C/C++, thereby having the characteristics of high test precision and high automaticity, and being applied to detecting the data race caused bytwo threads or multiple threads.

The invention relates to a multi-frame range image automatic registration method. (1) Any two frames of range images are registered through using SIFT characteristics, and the correctness of a result is judged. First, the SIFT characteristics of the two frames of the range images are calculated, corresponding points are bidirectionally crossly matched, then polar line constraint is calculated by a RANSAC algorithm, error matching is filtered, then the two frames of the range images are accurately registered through using an ICP algorithm, and the correctness of a result is judged. (2) a cycle space of a model diagram is researched, and a global consistent registration result is calculated. First, a derived cycle basis of the model diagram is calculated, an adjacency relational graph of the derived cycle basis is built, then a group basis of the consistent cycle space is calculated, and further a consistent registration result is obtained. The method can effectively increase the research speed of the cycle space, and an exponential time complexity can be increased to a line time complexity in an ideal condition. (3) the consistency of a cycle is judged, and the consistency of the registration result in the cycle is judged by a judgment method of relative differences. The invention can reliably automatically register multi-frame range images, through researching consistent cycles, error registration is taken out, thereby obtaining consistent registration results.

The invention discloses a message queue-based application software response time measuring method. The method comprises the following steps of: calling an ARM application interface library which sends a message to a message queue while application software calls function; reading the message in the message queue and calculating the function call response time of the application software; saving a function calling relationship, counting response time and saving a statistical result in a database; and displaying a relational graph between the statistical result and function call. A message queue-based application software response time measuring system comprises an ARM application interface library, a data acquisition agent program module, a data processing module and a data display module. The method has the advantages of helping maintenance personnel analyze and apply a running situation more completely, monitoring the complex mutual calling relationship of a service system, understanding abnormity causes and the like of service response time and perfecting a centralized monitoring system. Simultaneously, a failure analysis tool can be provided for a developer in a test developing process.

A method of perceptual 3-dimensional (3D) shape description and a method and apparatus for searching a perceptual 3D graphics model database established using the description method are provided. The description method includes: generating nodes that respectively correspond to parts of a part-based representation of a 3D shape model, the nodes including unary attributes of the parts; generating edges that include relational attributes between the nodes; and generating an attributed relational graph of the 3D shape model that is comprised of the nodes and the edges. The search method includes: receiving a predetermined 3D graphics model; transforming the received 3D graphics model into a perceptual 3D shape descriptor; and comparing the perceptual 3D shape descriptor with each of the perceptual 3D graphics models stored in the database to retrieve the 3D graphic models that are similar to the perceptual 3D shape descriptor. The searching apparatus includes: a query input unit that receives a query that is a 3D graphics model; a model/shape descriptor transforming unit that transforms the 3D graphic model received as the query into a perceptual 3D shape descriptor; a matching unit that compares the perceptual 3D shape descriptor with each of the perceptual 3D graphics models stored in the database to retrieve the models that are similar to the perceptual 3D shape descriptor; and a model output unit that outputs the retrieved model. A query by sketch or a query by editing is available, and the models that are similar to a query can be more accurately retrieved due to a double earth mover's distance method used to match query and model graphs.

The invention discloses a method and a system for individually recommending network commodities. The method includes firstly, acquiring purchasing behavior historical data of users and creating user-commodity purchasing relational graphs on the basis of the purchasing behavior historical data; secondly, adding a virtual user who purchases all commodities in the past into each user-commodity purchasing relational graph; thirdly, describing preference degrees of the specific target recommendation users to the commodities and giving new energy values to all the commodities; fourthly, ranking the commodities which are not purchased by the users according to a high-low sequence of the energy values of the commodities to generate individual recommendation lists. Each list can contain one or a plurality of commodities.

The embodiment of the invention discloses an entropy operation-based network intrusion detection method, which comprises the following steps of: capturing a network node data packet, and preprocessing the network node data packet to obtain target data; constructing a relational graph by utilizing the target data; calculating cross entropies of all network nodes; and sorting the cross entropies of all the network nodes, and finding out key network nodes with high activity. The embodiment of the invention also discloses an entropy operation-based network intrusion detection device. According to the entropy operation-based network intrusion detection method and the entropy operation-based network intrusion detection device disclosed by the invention, a network structure is converted into a graph structure, and the influences of the network nodes in the graph structure are found out according to the attributive characters of the network nodes in the graph by utilizing the entropy theory of the graph, and sorting is performed by the influences, and thus, the key network nodes with the highest activity can be easily obtained, and thereby, the information of the network nodes can be conveniently further analyzed so as to determine whether a network intrusion action occurs or not or adopt corresponding measure.

The invention discloses a type extracting system for a professional literature intellectual entity. The system comprises a user inquiring and feedback interface, an online crawler and management module, an intellectual entity identification module, an intellectual entity type extracting module, a type label propagation and index database establishment module, an intellectual entity type relational graph model establishment module, and a data visualization module. The system provided by the invention can be used for extracting the type of the entity according to key words of the entity searched by a user, and presenting a type relation and a hierarchical relation between intellectual entities and a timing sequence evolution mode in a visualization manner. Furthermore, the invention further provides a type extracting method for the professional literature intellectual entity. With the adoption of the method disclosed by the invention, type label extraction can be effectively carried out on the intellectual entities in professional fields; the problem of limitation and subjectivity in manual type predefinition can be solved; and the structured realization of a professional knowledge network is easily realized.

The invention discloses an establishment method of enterprise knowledge map based on a graph model. The method comprises the following steps that model establishment step: a graph data model is built,the graph data model comprises a graph data structure module, a graph mining algorithm module and a query interface module; data record step: each dimension data of a sample enterprise is obtained, according to the graph data model, dimension data such as shareholders, external investment, public opinion, judicial and the like of the enterprise is recorded in the graph data structure module; encapsulation step: according to common application scenarios, the graph mining algorithm is encapsulated into the query interface module, and a relational graph model is returned in a JSON format. The invention further discloses a graphical query method based on the graph model.

The invention discloses a method for detecting the tiny leakage of an oil transmission pipeline. The method comprises the following steps of: arranging a strong magnet at a mileage peg along a pipeline; transmitting a spherical inner detector into the pipeline at the head end of a pipe section to be detected, and getting the spherical inner detector back at the tail end of the pipe section to be detected; transmitting sound, accelerated speed, magnetism and time data which are recorded by a memory in the spherical inner detector to a computer; processing a magnetic signal, a sound signal and an accelerated speed signal respectively to obtain a relational graph of amplitude values and time of the signals; extracting a sound abnormal signal to determine the time when leakage points occur; and determining distance between the leakage points and a certain mileage peg according to the perimeter of the spherical inner detector and periodicity of the accelerated speed signal between the time of the leakage points and the time of a certain magnetic spike signal. The method has the advantages of high detection sensitivity, difficulty in blockage, convenience in use and accurate positioning.

The invention provides a piecewise linearity decoupling method based on a kalman filter and repeated collection, and aims at solving the problem of coupling between dimensions of a multivariate force sensor. The linearity decoupling method includes the steps of building a coupling error model, respectively deducing a force value-voltage input-output formula to each force value of the multivariate force sensor in the positive and negative directions, then carrying out a static calibration experiment on the multivariate force sensor, carrying out filtering on voltage output by the multivariate force sensor according to a kalman filter method, carrying out repeated measurement on each loaded force value to obtain an input-output relational graph between the loaded force values and the voltage output by the multivariate force sensor, then obtaining undetermined coefficient of a force value-voltage fitting formula according to the method of solving least squares solution of overdetermined linear equations, and obtaining a coupling matrix by carrying out coefficient matrix inversion to finish the coupling process. Compared with a traditional coupling method, the piecewise linearity decoupling method can remarkably improve the coupling accuracy of the multivariate force sensor on the premise of not increasing calculated amount.

The invention relates to a gear shifting method using economy shift schedule of a blade electric vehicle. According to the method, firstly the optimal economy shift schedule curve is made, and driving characteristic curves and efficiency relational graphs of a motor in 1 gear and 2 gear are made in the same graph; a horizontal line of constant driving force torque T0 is made in an overlapped area of the efficiency graphs in the 1 gear and the 2 gear; efficiency value curves, corresponding to different vehicle speeds in the 1 gear and the 2 gear, of traction T0 are made in a graph, and the intersection of the two curves is an ideal gear shifting point; opening degree values of an accelerator of upshift points and downshift points of the 1 gear and the 2 gear are calculated; the obtained upshift points and the downshift points are linked, so the optimal economy shift schedule can be obtained. Then the shift schedule curve is written in a TCU gear shifting control program. A TCU of the electric vehicle receives an accelerator opening degree signal and a vehicle speed signal of a driver in gear shifting, after judgment, when an ideal gear shifting point reaches, the TCU controls a gearbox gear shifting actuator, and thus gear-shifting motion is achieved.

A video summarized method based on mining the story structure and semantic relations among concept entities has steps of processing a video to generate multiple important shots that are annotated with respective keywords: Performing a concept expansion process by using the keywords to create expansion trees for the annotated shots; rearranging the keywords of the expansion trees and classifying to calculate relations thereof; applying a graph entropy algorithm to determine significant shots and edges interconnected with the shots. Based on the determined result of the graph entropy algorithm, a structured relational graph is built to display the significant shots and edges thereof. Consequently, users can more rapidly browse the content of a video and comprehend if different shots are related.

The invention relates to a massive email analyzing method and a system based on a relational graph. A massive email analyzing method based on the relational graph comprises the following steps: parallelly analyzing email source data, extracting head information and text information of an email and then storing the head information and text information of the email to an email list; storing parallelly analyzed summary information of an attachment on an email attachment list in a setting structure, and conducting detecting; constructing an email relational chart according to analyzed email data, and generating a single-point or multi-point relational graph according to user need and the email relational chart; introducing internet protocol (IP) address geography information database and email user identity information database, conducting relational analyzing on the email list, and displaying relational information on the generated relational graph. A massive email analyzing system based on the relational graph correspondingly comprises a parallel analyzing module, an attachment storing detecting module, a relational graph generating module and a relational analyzing module. The massive email analyzing method and the system based on the relational graph can effectively solve problems of massive email analyzing and processing and spam tracing and positioning in email network.

The invention discloses a method and a system for predicting power of a microgrid group. The method comprises steps: a microgrid graph theory model is built according to a communication relation between sub microgrids in the microgrid group; according to an intelligent agent node consistency theory and in combination with a micro Calman filter prediction algorithm, a power prediction algorithm for distributed prediction for needed power by a total grid of the microgrid group is determined; on the basis of the already-built communication relation graph theory topology, each sub microgrid in the microgrid group integrates grid needed power information of a local sub microgrid system and an adjacent sub microgrid system and uses a prediction algorithm to distributively perform short-term prediction on the needed power by the total grid of the microgrid group in the local sub microgrid system. Short-term prediction is carried out on the overall grid needed power of the microgrid group formed by multiple microgrids, and strong guarantee is provided for aspects of stable operation and power dispatching for the microgrid group power sharing system.

The invention discloses a traffic generation amount prediction system in a seaport planning year based on cargo back stepping, the technical scheme comprises the following steps: 1, a relational graph of a traffic generation amount and various traffic characteristic parameters and a traffic amount flow chart which is led by various cargo throughput are built, and current situation traffic characteristic parameters which are needed are collected. 2, the current situation traffic generation amount of a container truck, a common freight car and a bus is predicted. 3, the predicted current situation traffic generation amount is compared with the traffic generation amount which is obtained based on a current situation traffic investigation. 4, above related parameters are calibrated to obtain a group of best parameter values through a genetic algorithm. 5, prospect factors are comprehensively considered, the current situation traffic characteristic parameters of various cargo are amended, thereby obtaining the traffic characteristic parameters in a planning year. 6, various cargo public road transportation volume, the traffic characteristic parameters in the planning year and the current situation in the planning year of a building area are considered, and the prediction of the traffic generation amount in the planning year is completed based on cargo back stepping traffic amount.

The invention relates to a static detection method and apparatus for a repackaged malicious application. The method comprises the steps of obtaining API call sequences of an installation package of a to-be-detected application and a correlation relationship between classes which the API call sequences belong to; establishing a class-based function call relational graph; performing clustering division on the classes to obtain a plurality of clusters, and removing a preset number of clusters with the highest correlation relationship between the classes in the clusters to obtain a malicious code cluster; extracting sensitive API call sequences in the API call sequences of the classes in the malicious code cluster, and performing similarity matching on the extracted sensitive API call sequences of the classes and characteristic sequence samples of the malicious application in a pre-established sample library; and determining whether the to-be-detected application is the repackaged malicious application or not. According to the method and the apparatus, malicious codes are extracted by taking the class as a unit without depending on an Android official application, so that relatively high accuracy can be ensured for variant malicious codes.

The invention relates to the fields of deep learning, natural language processing and the like, and particularly relates to a syntactic relationship enhanced machine reading understanding multi-hop reasoning model and method. The model comprises a text coding module, an association element relationship graph construction module, a multi-hop reasoning module, an answer generation module and an answer prediction module. According to the invention, syntactic relationships are fused into the graph construction process, an associated element relationship graph is constructed, multi-hop reasoning iscarried out by utilizing a graph attention network based on the relationship graph, and answer support sentences are mined; meanwhile, a multi-head self-attention mechanism is introduced to further mine text clues of viewpoint type questions in the article, and an automatic solution method of the viewpoint type questions is improved; and finally, a plurality of tasks are subjected to joint optimization learning, so that when the model answers the questions, fact description for supporting the answers can be given, the interpretability of the model is improved, and meanwhile, the existing answering method for viewpoint type questions is improved.

The invention discloses an unsupervised transfer learning method based on a graph convolution network. The method comprises the steps of obtaining source domain and target domain samples for transferlearning from a database, performing feature extraction on the samples, and constructing a correlation graph; putting the sample features and the relational graph into a constructed graph convolutional network, mapping the sample features to a feature space with strong discrimination, and forming new features of the sample; performing distribution alignment on the learned feature space and the newsample features, so that the new sample features have good migration performance; meanwhile, constructing a classification network and learning a classifier of target domain data; and repeatedly using the gradient descent method until the loss functions of the graph convolution network and the classification network converge, and predicting the unlabeled target domain data. The method combines two characteristics of model discrimination capability and knowledge migration capability, can be used for a difficult unsupervised migration learning scene, and has good classification learning and data labeling capability in the application of an actual scene.

The invention discloses a display method and device of a relational graph. The method includes: acquiring a list of paths between a first target node and at least one second target node, wherein the list of the paths can include n connection paths, and the n connection paths have m associated nodes; carrying out primary sorting on the n connection paths according to appearance frequency of the nodes, and carrying out secondary sorting on the n connection paths according to similarity degrees of the connection paths; for each connection path after the secondary sorting, respectively calculating coordinates of each associated node, which is included in each connection path, according to size information of a target display interface; merging coordinates of common nodes to obtain final coordinates of each common node; and drawing all the nodes on the target display interface according to the coordinates thereof, and drawing connection lines. Therefore, locations of the nodes on the target display interface can be effectively distributed, and a neat display can be formed.

The invention discloses a social network based mobile terminal user grouping method. The method comprises: according to history of communication between terminal users, quantizing communication contact to generate a social relational graph (STG); in combination with preference attributes of the terminal users, generating an attribute relational graph (ARG) taking preference degrees between the terminal users and attributes as weights; generating a social relation-attribute graph in combination with the STG and the ARG, designing an SAPLA algorithm to predict unknown attributes of the terminal users, and adjusting preference degrees of known attributes; and proposing an SARA algorithm by utilizing a random walk model, combining transfer probabilities between the terminal users and between the terminal users and the attributes, giving out a transfer probability matrix between the terminal users, with relatively low complexity, giving out a random walk distance matrix Rl by utilizing the transfer probability matrix, setting a target function in combination with the matrix Rl, and grouping the terminal users until the target function is converged. According to the method, the complexity of operation is lowered and the accuracy of grouping is improved.