117 results about "Machine learning" patented technology

A method and system for fingerprint template synthesis from multiple fingerprint images is provided. A first set of minutiae points is extracted from a first fingerprint image. A second set of minutiae points is extracted from a second fingerprint image. The orientation is calculated for a plurality of minutiae points selected from the first set of minutiae points based on the first fingerprint image. Simulated points are added to the first set of minutiae points, wherein simulated points are created based on the location and orientation of minutiae points in the plurality of minutiae points. The first set of minutiae points and the second set of minutiae points are registered and the first set of minutiae points and the second set of minutiae points are combined.

The invention relates to methods and systems for semantic disambiguation of a plurality of words. A representative method comprises providing a dataset of words associated by meaning into sets of synonyms; locating said sets at respective vertices of a graph according to semantic similarity and semantic relationship; transforming the graph into a Euclidean vector space comprising vectors indicative of respective locations of said sets; identifying a first group of said sets which include a first of said pair of words; identifying a second group of said sets which include a second of said pair of words; determining a closest pair in said vector space of said sets taken from said first and second groups of sets respectively; and outputting a meaning, of said plurality of words based on said closest pair of said sets and at least one of said semantic relationships between said closest pair of said sets.

The invention provides a synchronizer self-learning identification control method and a vehicle synchronizer position verification control method based on the synchronizer self-learning identification control method. According to the synchronizer self-learning identification control method, forward and reverse rotation of a gear selection and shifting motor are controlled, the voltage change of a gear selection and shifting position sensor is monitored so as to identify the state of a synchronizer reaching extreme positions, the voltage value of the position sensor when the synchronizer is at each extreme position is recorded, and therefore self-learning identification of gear selection information and gear shifting information of the synchronizer is realized. According to the position verification control method, when a speed changer reaches a bottom line, a vehicle runs for a certain period of time and the vehicle cannot be started normally due to gear selection and shifting, the synchronizer self-learning identification control method is adopted for learning and identification of the gear selection information and the gear shifting information. By means of the synchronizer self-learning identification control method and the position verification control method, the problem that the synchronizer is inaccurate in position because of multiple factors is effectively eliminated, gear selection and shifting of the vehicle can be based on a correct synchronizer position, and therefore the correctness of gear selection and shifting of the vehicle is guaranteed.

The invention belongs to the technical field of tunnel engineering, and discloses a dynamic risk acquisition method for tunnel large deformation disasters, which comprises the following steps: S1, identifying risk disaster-causing factors according to historical data of the tunnel large deformation disasters; S2, selecting a prediction index according to the risk disaster-causing factor, and establishing a probability prediction model according to the prediction index; S3, dynamically updating surrounding rock parameters by using a Bayesian method and a Markov random process method according to the prediction index and the exposed surrounding rock information of tunnel face excavation; S4, according to the surrounding rock parameters, the prediction indexes and the probability prediction model, carrying out posterior risk updating and acquiring a dynamic risk prediction result of the tunnel large deformation disaster. The method solves the problems that in the prior art, a set of quantitative risk evaluation model capable of reflecting a large deformation action mechanism cannot be established, effective feedback of surrounding rock and support information in the construction period cannot be effectively utilized, and a dynamic and informationized risk evaluation theory cannot be formed.

The invention discloses a method for detecting a high risk of a bankcard. The method comprises steps of: clustering the historical data of each bankcard transaction by using a k-means clustering algorithm in order to obtain a risk model, wherein the historical data, used as training data of the risk model, is classified into a high-end card type and non-high-end card type, and each type of information data is represented by n dimensions; processing the transaction data of a bankcard to be detected into data with a dimensionality the same as that of the training data of the risk model; determining, according to the risk model, whether the processed data complies with a rule and characteristic changing from the non-high-end card type to the high-end card type; and if not, determining that the bankcard is provided with the high risk. The invention also discloses a data processing device.

The invention discloses a community question and answer website credible evaluation method based on a network, comprising the following steps: 1) constructing an answer-user association network; 2) based on the construction of the answer-user association network, using the mutual inference algorithm to acquire the user credibility and the credibility of the answer in an iterative manner., and themethod can evaluate the credibility of the answer of the community question answering website.

The invention discloses a human body analysis method based on graph representation and an improved Transform. According to the method, high-dimensional feature representation is embedded into low-dimensional graph features, the improved Transform is used for reasoning calculation and capturing context feature relations, new graph features are generated and decoded again to form a fine analysis graph, and therefore the whole model is iteratively trained in an efficient mode to obtain a final analysis result. According to the method, reasoning calculation is performed more efficiently only according to priori knowledge of a human body hierarchical structure; reasoning is performed on the human body part features represented by the graph, so that more calculation cost can be saved in subsequent iterative reasoning; and the structure of Transform is improved, and the context information of the features of each part of the human body is extracted and integrated globally, so that the association degree of different parts of the human body is perceived comprehensively, and the precision of an analysis result is higher.

The invention relates to the technical field of text recognition, in particular to a knowledge graph hyponymy relation recognition method based on pattern expansion and BERT classification and a storage device. The knowledge graph hyponymy relation recognition method based on pattern expansion and BERT classification comprises the steps: extracting potential hyponymy relation pairs from external data based on pattern expansion, and forming a hyponymy relation model training seed corpus in combination with a preset resource library; obtaining a to-be-predicted hyponymy data set, and predictingthe to-be-predicted data set based on the BERT-Attention-Bi-LSTM model to obtain a hyponymy prediction result; and further processing the hyponymy relationship prediction result through a preset ruleto obtain a final hyponymy relationship prediction result. The method greatly reduces the complexity and labor investment of rule compilation, has stronger realizability compared with another mainstream statistics-based hyponymy concept identification method, and can provide technical support for the construction of various professional knowledge maps.

The invention provides a method and a device for generating an automatic identification fall model. The method comprises the following steps: acquiring a fall data set; identifying the fall data by using a historical identification model to obtain fall data with error and missing identification; preprocessing the data with typical characteristics in the tumble data and the tumble data with wrong identification and missing identification to obtain preprocessed training sample data; convolution operation and normalization processing of a neural network are carried out on the preprocessed training sample data to obtain a detection result for identifying fall; and determining a final recognition tumble model according to a detection result. According to the invention, massive data is trained;according to the method, behaviors such as falling can be automatically recognized without participation of people, falling of old people can be found in time and rescued, damage to the old people isreduced to the minimum, the accuracy of the obtained final recognition falling model is usually higher than that of manual work, manpower input can be effectively saved, and the safety monitoring precision is improved.