604 results about "Distance matrix" patented technology

The Computationally Efficient Transfer Processing and Auditing Apparatuses, Methods and Systems (“CETPA”) transforms transaction record inputs via CETPA components into matrix and list tuple outputs for computationally efficient auditing. A blockchain transaction data auditing apparatus comprises a blockchain recordation component, a matrix Conversion component, and a bloom filter component. The blockchain recordation component receives a plurality of transaction records for each of a plurality of transactions, each transaction record comprising a source address, a destination address, a transaction amount and a timestamp of a transaction; the source address comprising a source wallet address corresponding to a source digital wallet, and the destination address comprising a destination wallet address corresponding to a destination virtual currency wallet; verifies that the transaction amount is available in the source virtual currency wallet; and when the transaction amount is available, cryptographically records the transaction in a blockchain comprising a plurality of hashes of transaction records. The Bloom Filter component receives the source address and the destination address, hashes the source address using a Bloom Filter to generate a source wallet address, and hashes the destination address using the Bloom Filter to generate a destination wallet address. The Matrix Conversion component adds the source wallet address as a first row and a column entry to a stored distance matrix representing the plurality of transactions, adds the destination wallet address as a second row and column entry to the stored distance matrix representing the plurality of transactions, adds the transaction amount and the timestamp as an entry to the row corresponding to the source wallet address and the column corresponding to the destination wallet address; and generate a list representation of the matrix, where each entry in the list comprises a tuple having the source wallet address, the destination wallet address, the transaction amount and the timestamp.

The present invention provides a method and system for positioning one or more anchor nodes or one or more non-anchor nodes in one or more communication networks. A non-anchor node may be in communication with two or more anchor nodes. The method comprises determining two or more distance measurements, corresponding to raw distances of the non-anchor node from the two or more anchor nodes, at predetermined intervals of time. An estimated distance is then calculated between the two or more anchor nodes based on the two or more distance measurements. Estimated distances between a plurality of anchor nodes in the one or more communication networks is calculated in a similar manner. A partial distance matrix is then populated using the estimated distance between the plurality of anchor nodes. A plurality of geographic coordinates of the plurality of anchor nodes is reconstructed based on the partial distance matrix. The plurality of geographic coordinates is reconstructed based on one or more of a geometric build-up algorithm, a shortest-path algorithm and a multidimensional scaling algorithm.

The embodiment of the invention provides a multi-target tracking method and system based on depth features. The method comprises the following steps: obtaining detection frame positions correspondingto targets detected in a current frame image and the depth features of the targets; based on the position of the detection frame corresponding to each target in the previous frame of image, obtainingthe prediction position of each target in the current frame by using a Kalman filter; according to the detection frame position corresponding to each target, the prediction position of each target inthe current frame, the depth feature of each target and the depth feature set of each tracker, performing cascade matching on the detection frame corresponding to each target and the tracker by usinga Hungarian algorithm; And calculating an IOU distance matrix between the detection frame on the non-cascade matching and the tracker to be matched, and performing IOU matching between the detection frame and the tracker by using a Hungarian algorithm based on the IOU distance matrix to obtain a final matching set. According to the embodiment of the invention, the target tracking effect under theshielding condition can be effectively improved, and the number of times of ID switching is reduced.

A method for comparing data sequences includes accepting first and second data sequences of data elements. A distance matrix is computed. The matrix includes rows and columns of matrix elements, describing distances between the data elements of the first sequence and the data elements of the second data sequence. The distance matrix is reshaped by applying successive, incremental shifts to the rows or columns so as to produce a reshaped matrix. A best-score path through the reshaped matrix is calculated using vector operations, so as to quantify a similarity between the first and second data sequences. Due to vectorization, a significant increase in computation speed is achieved in both software and hardware implementations.

In accordance with the teachings described herein, systems and methods are provided for analyzing transactional data. A similarity analysis program may be used that receives time-series data relating to transactions of an organization and performs a similarity analysis of the time-series data to generate a similarity matrix. A data reduction program may be used that receives the time-series data and performs one or more dimension reduction operations on the time-series data to generate reduced time-series data. A distance analysis program may be used that performs a distance analysis using the similarity matrix and the reduced time-series data to generate a distance matrix. A data analysis program may be used that performs a data analysis operation, such as a data mining operation, using the distance matrix to generate a data mining analysis of the transactional data.

A method discovers patterns in unknown multi-dimensional data. A time-series of the multi-dimensional data is generated and a point cross-distance matrix is constructed by self-correlating the time-series. All minimum cost paths in the point cross-distance matrix are located at multiple time resolutions. The minimum cost paths are then related to temporal sub-sequences in the multi-dimensional data to discover high-level patterns in the unknown multi-dimensional data.

The invention discloses a multiple inspection robot cooperative operation method for a substation sequence control system. When the closed-loop control and the state visualization of the intelligent substation sequence control system are implemented, the executing and checking time of the sequence control system can be effectively saved by adopting the method. The invention provides the multiple inspection robot cooperative operation method which comprises the following steps: modeling by adopting a topological method on the basis of the substation environment, forming a topological network chart and a communication relation matrix according to walking routes of inspection robots, and carrying out search and traversal of paths to output a discharge matrix and a shortest path matrix; in the single-step sequence control process, determining a shortest communication path sequence of each inspection robot from a current position to a target point according to the distance matrix and the shortest path matrix; carrying out excellence selection of single-step sequence control, i.e. selecting the inspection robot with the optimal performance under the condition of considering comprehensive indexes of journey, time, safety and maneuverability; determining an operation path sequence of each inspection robot and a corresponding activation event; and carrying out on-line monitoring, wherein if the environment is changed, the on-line monitoring takes charge of debugging and recovering operation.

The invention provides a sensor network node self-positioning method based on the relation of distance, size and sequence, belonging to the field of wireless sensor network self-positioning technique. The method of the invention is characterized in that the method sequentially comprises the steps as follows: node establishing, routine establishing and data transmission, network connecting information extracting, node relative coordinate positioning, absolute coordinate converting, compute result outputting and the like. When the relative coordinate of the node is being positioned, the relation of distance, size and sequence of the shortest path distance is used for obtaining the initial estimation topology of the node position; based on the initial estimation topology, a logarithm likelihood function of optimum node distance matrix and inverse solution Euclidean distance used as initial estimation topology; subsequently, an adjacent module comparison method is used to alternatively iterate the optimum distance matrix and the Euclidean distance; the method of the invention is not dependent on the detailed measurement value of the distance, only uses the relative sequence relation of the distances between nodes and has unique advantages.

The invention discloses a target identification method based on contour features. The method comprises steps of establishing a feature bank of object template contours; extracting a complete contour of an object template; extracting feature points and central points from the contour; establishing a distance matrix to describe the contour by using feature points and central points; conducting calculation of the distance matrix by aiming to all pixels on the contour; conducting target identification for an image to be detected; extracting an edge of the image to be detected; extracting feature points on the edge; calculating feature description composed of feature points; matching the feature description of the image to be detected with features in the feature bank of object template contours; estimating central points of the contour of the image to be detected; and estimating the contour of the image to be detected. Compared with the prior art, the method has the advantages that the size problem in the contour matching is solved, and the size in the contour matching is not changed. The method can be applied to target identification of images effectively.

The present invention relates to a shortest path-based automatic positioning method of sensor network nodes on the sparse measurement set, and belongs to the technical field of automatic positioning of the wireless sensor network. The positioning method is characterized by orderly comprising node distribution, route establishment, data transmission, network connection information extraction, relative coordinate positioning of the node, absolute coordinate transformation, outward transmission of results by a computer, and so on; wherein, when the relative coordinates of the node are positioned relative to the adjacent node, a detection method of radio frequency is used for measuring the distance between the nodes; for the non-adjacent node, the distance of Freud shortest-path is used for estimating the distance between the nodes, thus the shortest-path distance matrix which comprises the adjacent node and the non-adjacent node is got; then a method of multi-dimensional scaling analysis can be adopted to acquire the initial value of topological estimation of the node position which is relatively positioned; the similar probability distribution of the unknown distance is calculated; the likelihood function of the complete-distance matrix is then used as an expected objective function for optimization, so as to eliminate the randomness, thus the relative positioning results can be acquired.

The invention provides a wireless indoor positioning method. The method comprises the following steps of: by the adoption of an intelligent cellphone, automatically collecting wireless fingerprint data and user mobile data, forming a fingerprint collection F and a distance matrix D', and carrying out pre-treatment; according to the fingerprint collection F and the distance matrix D' after pre-treatment, building a fingerprint space; and generating a plane graph without stress, carrying out extraction of key characteristics, and carrying out space coordinate conversion, wherein the extraction of key characteristics comprises recognition of a passageway, recognition of a room and matching of a reference point, and the space coordinate conversion comprises conversion of a floor level and conversion of a room level. According to the wireless indoor positioning method, a manual site survey is not needed for a positioning area, the fingerprint information can be conveniently provided by coordination of multiple users in a system, the precision on a positioning result is high, the logic is excellent, and redundant information generated in inquiry and usual movement can be taken as update information.

Systems and methods for clustering content according to similarity are provided that identify and group similar content using a set of tags associated with the content. A topic model of a group of content is built, producing a probability distribution of topic membership for the content. Individual items of content are then clustered using a clustering algorithm, and a distance matrix from the probability distribution is built. Based on the distance matrix, individual items of content are labeled as “must-link” or “cannot-link” pairs with the group of content. The topic model is then embedded into successively smaller dimensions using a kernel method, until the clustering is stable with respect to both the behavioral and content domains.

The present invention provides a re-recognition network training method, device and system, and a re-recognition method, device and system. The re-recognition network training method comprises the steps of: obtaining batch processing training data; obtaining feature vectors corresponding to N images included in the batch processing training data, calculating the distance between two each feature vectors according to the feature vector corresponding to each image, and obtaining a distance matrix; and according to the distance matrix obtained through calculation, selecting a positive sample pairwith the maximum distance and a negative sample pair with the minimum distance, employing the two selected boundary samples to calculate loss of a convolutional neural network to train a model. Therefore, the most difficult positive sample pair and the negative sample pair are learned to calculate the loss of the convolutional neural network so that the generalization capability of the convolutional neural network model can be improved and the identification precision is improved.

The present invention discloses a method for pedestrian weight recognition in a video surveillance scene. The method is implemented through the following steps: S1, pre-training an FT-FNN network; S2, fine-tuning the FT-FNN network; S3, extracting a depth feature and an attribute feature of a training image; S4, optimizing the attribute weight; S5, extracting the attribute feature of a to-be-recognized image; S6, extracting the attribute feature of a pedestrian bank; S7, generating a distance matrix; and S8, outputting matching images in sequence. The depth feature after fine-tuning has strong distinguishing ability for the pedestrian weight recognition; compared with the texture, color and other low-level visual features, the middle-level semantic attributes are stable and not easy to change greatly with the change of the light and the posture; and manual integration of low-level vision features can improve the distinguishing ability of the depth feature to a certain extent, and can improve accuracy of some middle-level attributes closely related to the color and the texture, so that by combining the three advantages, the method disclosed by the present invention achieves better accuracy in the field of pedestrian weight recognition.

Techniques are presented for analyzing audio-video segments, usually from multiple sources. A combined similarity measure is determined from text similarities and video similarities. The text and video similarities measure similarity between audio-video scenes for text and video, respectively. The combined similarity measure is then used to determine similar scenes in the audio-video segments. When the audio-video segments are from multiple audio-video sources, the similar scenes are common scenes in the audio-video segments. Similarities may be converted to or measured by distance. Distance matrices may be determined by using the similarity matrices. The text and video distance matrices are normalized before the combined similarity matrix is determined. Clustering is performed using distance values determined from the combined similarity matrix. Resulting clusters are examined and a cluster is considered to represent a common scene between two or more different audio-video segments when scenes in the cluster are similar.

Large amounts of multimedia data are transmitted over information networks in the form of a digital stream, analog video, or text captioning. Often, repetitions such as paid advertisements, theme music at the commencement of a TV broadcast, and common jingles and slogans occur in these streams. Detection of repetitions in a transmitted signal such as streaming audio or video is described, and includes extracting a plurality of samples from the information stream and accumulating the samples into segments comprising an interval of the transmitted signal. A vector indicative of the samples in each of the segments is generated, and each of the vectors in the segments is correlated to generate a covariance matrix, or signature, corresponding to the segment. Each of the covariance matrices are aggregated into a sequence of covariance matrices and compared to other covariance matrices to generate a distance matrix. The distance matrix includes a distance value, indicative of the similarity between the distance matrices, as a result of the comparing of each matrix. The distance matrix is then traversed to determine similar sequences of covariance matrices.

The invention discloses a method and a device for route optimization of logistics delivery vehicle, and belongs to the technical field of logistics. The method comprises the following steps of: initializing a congestion matrix alpha and a distance matrix D, generating a delivery route weight matrix omega=alpha D, and initializing a population module N<ZQ>; selecting a population size N<X>, a maximum number of generations N<G>, a crossing-over rate beta, a mutation rate gamma and a number of generations n=0, generating an initial route r1 through a greedy algorithm, and performing mutation operation on the initial route r1 to generate N<ZQ>-1 new routes; calculating fitness A<n> of each route of a first generation population formed by the initial route and the new routes, selecting N<X> routes with the highest fitness from the current population by adopting selection operators, and performing crossover and mutation operations on the N<X> routes to generate a population of next generation; updating n=n+1, when n=N<G>, calculating the fitness A<n> of all the routes in the latest population, and selecting the delivery route with the highest fitness in the current population as the optimal route. According to the invention, when the logistics delivery vehicle delivers goods, the delivery time can be as less as possible, and the delivery route can be as short as possible.

The invention belongs to the technical field of multisource information fusion, and discloses a flight path fusion method. The flight path fusion method comprises the following steps of: establishing a relative distance matrix between data by using observation information of a plurality of sensors; computing a support threshold function to obtain a support threshold matrix, establishing an equation set, and solving a weighting factor; multiplying the weighting factor with a corresponding observed value, obtaining corresponding filter values respectively through filtering, and adding all obtained filter values to obtain a filter fusion value with an observation coefficient; and updating an estimated value of target state step by step by using Kalman filtering, wherein the filter fusion value serves as a state updating input value. In the invention, as the filter fusion of the observation coefficient is carried out through the observation information of the plurality of sensors, influence on the fusion of flight paths due to the uncertainty of the observation information is reduced under the condition that data processing complexity is not increased; and correlation of the observation information is taken into consideration during the filter fusion of the observation coefficient, so that observation accuracy is increased, and the reliable tracking of a target is obtained.

The invention provides a kernel method-based collaborative filtering recommendation system and a kernel method-based collaborative filtering recommendation method. The corresponding system comprises a data preparation module which is used for standardizing the original data and carrying out corresponding preprocessing, generating a user-project rating matrix and a project distance matrix to output; a user interest modeling module which is used for constructing an interest model for a user on a project space according to the user-project rating matrix and the project distance matrix as well as a kernel density estimation technology; and a recommendation result generation module which is used for computing the similarities among the users according to the interest model, generating a neighbor set of a target user, and predicting a score of the project rated by the user according to a predetermined recommendation strategy and returning the recommendation result. Through the recommendation system and the recommendation method provided by the invention, the user interest model can be better presented, the user similarity in the practical application is estimated more accurately, the performance of the recommendation system can be promoted considerably, and more stable recommendation result can be obtained.

The invention relates to the communication field, and provides a cluster communication terminal track real time anomaly detection method and system based on hybrid grid hierarchical clustering. The method comprises the steps: the step 1: constructing a track based on grids, and determining the size of the optimal grid; the step 2: calculating a Hausdroff distance matrix, utilizing a Hausdroff distance formula to calculate the distance between all the tracks based on the tracks of the grids, and generating a distance matrix of a track set; the step 3: hierarchical clustering, that is, based on the Hausdroff distance matrix of the track set, applying an agglomerate hierarchical clustering algorithm from bottom to top to realize classification of normal and abnormal tracks of a large scale of tracks; and the step 4: anomaly detection method evaluation and feedback: utilizing the above method to perform anomaly track detection on the track set which has a track classification identifier to obtain an anomaly classification result, and evaluating whether a model parameter is reasonable after comparison and making a feedback. The cluster communication terminal track real time anomaly detection method based on hybrid grid hierarchical clustering can realize on-line real time detection of an anomalous event, and can improve the upper layer dispatching efficiency of a cluster communication system.

The invention discloses a ship navigation path planning method, system, medium and equipment, and the method comprises the steps: building an offshore logistics condition model, determining the numberof ports and ship carrying capacity of a ship navigation area according to the ship navigation plan, an offshore climate environment and the change condition of topography and geomorphology in a route, and determining the number of ships needing to be dispatched; constructing an objective function and a constraint condition of an optimal path for the paths of the ships needing to be dispatched among different ports; establishing a ship navigation path optimization model; combining historical data of a ship navigation plan; and calculating an optimal navigation path by adopting an ant colony algorithm: initializing defined parameters, converting port coordinate information into a distance matrix between ports, searching for paths between different ports by utilizing ants, recording iteration times and updating pheromone concentration according to path lengths, and outputting the optimal path after the maximum iteration times are reached. According to the invention, the economic cost consumed by offshore logistics and invalid driving of the ship are reduced, and the economic benefit is improved.

The embodiment of the invention discloses an error correction method and device for a voice recognition text. The error correction method for the voice recognition text comprises the steps that at least one candidate error correction text used for error correction of a result text is obtained according to the multi-layer K-Gram index of the voice recognition result text, the fuzzy sound editing distance matrix between the at least one candidate error correction text and the result text is determined, the fuzzy sound editing distance between the at least one candidate error correction text and the result text and a candidate error correction boundary are obtained according to the determined fuzzy sound editing distance matrix, an error correction text is selected according to the fuzzy sound editing distance corresponding to the at least one candidate error correction text, and error correction is conducted on the result text according to the candidate error correction boundary corresponding to the error correction text. By the adoption of the error correction method and device for the voice recognition text, accurate error correction of the voice recognition result text is achieved.

The invention relates to an apparatus for clustering process models each consisting of model elements comprising a text phrase which describes in a natural language a process activity according to a process modeling language grammar and a natural language grammar, wherein said apparatus comprises a process object ontology memory for storing a process object ontology, a distance calculation unit for calculating a distance matrix employing said processing modeling language grammar and said natural language grammar, wherein said distance matrix consists of distances each indicating a dissimilarity of a pair of said process models, and a clustering unit which partitions said process models into a set of clusters based on said calculated distance matrix.

A method for determining the geometry and / or the localisation of an object comprising the steps of:sending one or more signals by using one transmitter;receiving by one or more receivers the transmitted signals and the echoes of the transmitted signals as reflected by one or more reflective surfacesbuilding by a computing module a first Euclidean Distance Matrix (EDM) comprising the mutual positions of the receivers;adding to the EDM matrix a new row and a new column, the new row and a new column comprising time of arrivals of said echoes and computing its rank or distance to an EDM matrixdetermining the geometry and / or the position of the object based on said rank or distance.

The invention discloses an unsupervised cross-domain self-adaptive pedestrian re-identification method. The method comprises the following steps of S1, pre-training an initial model in a source domain; s2, extracting multi-granularity characteristics of a target domain by utilizing the initial model, generating multi-granularity characteristic grouping sets, and calculating a distance matrix for each grouping set; s3, performing clustering analysis on the distance matrix to generate intra-cluster points and noise points, and estimating hard pseudo tags of samples in the cluster; s4, accordingto a clustering result, estimating a soft pseudo label of each sample for processing noise points, and updating a data set; s5, retraining the model on the updated data set until the model converges;s6, circulating the steps 2-5 according to a preset number of iterations; s7, inputting the test set data into the model to extract multi-granularity features, and obtaining a final re-identificationresult according to the feature similarity; according to the method, the natural similarity of the target domain data is mined by utilizing the source domain and the target domain, the model accuracyis improved on the label-free target domain, and the dependence of the model on the label is reduced.

The invention relates to a mechanical vibration fault characteristic time domain blind extraction method, and belongs to the technical field of mechanical equipment status monitor and fault diagnosis. The mechanical vibration fault characteristic time domain blind extraction method includes: firstly, expanding a vibration observation signal into a high dimension signal subspace; then, obtaining a low dimension signal; afterwards, performing FastICA independent component analysis, calculating normalization kurtosis of all independent components, figuring out a component signal corresponding to the minimum normalization kurtosis, and using an orthogonal matching pursuit algorithm to reconstitute periodic signals; subsequently, removing the reconstituted periodic signal from each independent component, and then using an improved KL distance algorithm to calculate a distance matrix among the independent components after the periodic signals are removed from the independent components, and performing dynamic particle swarm clustering so as to obtain an estimation signal; finally, analyzing an envelope demodulation spectrum of the estimation signal, and performing fault diagnosis. The mechanical vibration fault characteristic time domain blind extraction method is suitable for processing a long convolution data problem, can effectively reduce influences from periodic ingredients on a blind separation result, and simultaneously can solve blind separation result order uncertainty problems, and finally achieves bearing fault characteristic extraction.

The invention discloses a graph convolutional neural network based urban traffic flow prediction method and a medium. The method comprises that original data is obtained; a distance matrix is generated according to latitude and longitude information corresponding to each node; a reachability matrix is calculated according to a mean value of the speed limit and the distance matrix; an initial traffic flow prediction model for predicting the traffic flow velocity is established, traffic flow velocity information and the reachability matrix are input to the initial traffic flow prediction model,so that the initial traffic flow prediction model outputs a traffic flow velocity predicted value according to the traffic flow velocity information and the reachability matrix; the initial traffic flow prediction model is trained to determine a final traffic flow prediction model; and traffic flow velocity information to be predicted and the reachability matrix to be predicted are input to the traffic flow prediction model, so that traffic flow in the future is predicted via the traffic flow prediction model. Thus, spatial characteristics of the urban traffic road network are effectively extracted, the traffic flow is predicted more accurately, and the prediction method is more universal and convenient to popularize.

The invention discloses an object volume calculating method based on Kinect. The method comprises the steps of (1), acquiring a depth image and a color image by means of Kinect; (2), calibrating a color camera of the Kinect; (3), setting an ROI area of the depth image, performing image segmentation by means of a foreground color image which comprises a measured object and a background color image of a measurement platform that does not comprise the measured object, and obtaining a binary image of the measured object; (4), converting a background depth image ROI area to a background distance matrix, and performing preprocessing on the background distance matrix, filling in elements which are zero on the background distance matrix, and converting the foreground depth image ROI area to a foreground distance matrix; (5), obtaining a height matrix according to difference between the foreground distance matrix and the background distance matrix; and (6), calculating length, width, height and volume of the object. The object volume calculating method effectively settles problems of high labor intensity and long measurement time in traditional manual measurement, and is a noncontact measurement means. Damage of the measurement object is prevented. A requirement for automatic is satisfied and furthermore measurement precision is improved.