177results about How to "Low similarity" patented technology

The invention provides a radio frequency fingerprint positioning method and system as well as an access controller (AC) and a positioning server. The method comprises the following steps: the positioning server sends a first positioning indication message which contains an identifier of the appointed equipment to an access point (AP) through the AC, so the AP can collect the received signal strength indication (RSSI) of the appointed equipment; the AC receives the RSSI of the appointed equipment fed back from the AP, filters the RSSI of the appointed equipment according to a preset dimensional threshold value and an RSSI threshold value to generate a fingerprint feature vector at least containing the preset dimensional threshold value number of the RSSI, and sends the fingerprint feature vector to the positioning server; and the positioning server positions the appointed equipment according to the fingerprint feature vector and a preserved standard fingerprint feature vector containing at least the preset dimensional threshold value number of the RSSI and position coordinates. By utilizing the technical scheme of the invention, the complexity of positioning by using a radio frequency fingerprint technique can be decreased, and the positioning precision can be increased.

A user interface (UI) for adaptive video fast forward provides a novel fully adaptive content-based UI for allowing user interaction with an image sequence or video relative to a user identified query sample. This query sample is drawn either from an image sequence being searched or from another image sequence entirely. The user interaction offered by the UI includes providing a user with computationally efficient searching, browsing and retrieval of one or more objects, frames or sequences of interest in video or image sequences, as well as automatic content-based variable-speed playback based on a computed similarity to the query sample. In addition, the UI also provides the capability to search for image frames or sequences that are dissimilar to the query sample, thereby allowing the user to quickly locate unusual or different activity within an image sequence.

Disclosed are methods and systems for enabling analysis of communication content while preserving confidentiality. In one embodiment, communication content is processed to increase the similarity of superficially dissimilar instances of communication content and / or to increase the distinctiveness of superficially similar instances of communications content. In this embodiment at least part of the processed communication content is hashed to obscure the actual communication content. In one embodiment, social network analysis is performed on the communication content after hashing, and visualization of the social network analysis includes thread graphs and / or circular graphs.

The invention discloses a data subspace clustering method based on multiple view angles, which comprises the steps of extracting multi-view-angle characteristics in a multi-view-angle database; for the multi-view-angle database, selecting a specific linear reconstruction expression method and determining a regularization constraint method corresponding to the linear reconstruction expression method; determining reconstruction error weight of each view angle characteristic in multi-view-angle characteristics; according to the selected reconstruction expression method and the obtained reconstruction error weights of different view angle characteristics, learning to obtain a linear expression matrix for reconstructing all samples in the multi-view-angle database, wherein the linear expression matrices are used for expressing a relationship among the samples in the database and element values are used for expressing reconstruction coefficients for corresponding samples in the line to reconstruct corresponding samples in the row; correspondingly processing the linear expression matrix to obtain an affinity matrix for measuring the similarity of the samples in the multi-view-angle database; and using a spectral clustering algorithm to partition the affinity matrix to obtain multi-view-angle data subspaces.

The invention relates to a cancer predicting method and a drug design method. Specifically, the invention relates to a method for predicting cancer which is useful for the genetic diagnosis for evaluating the malignancy of cancer. The invention also relates to a method for designing a drug based on the result of the above prediction method.

An information retrieval system comprises: a speech input unit for inputting speech; an information storage unit for storing information with which speech information, of a length with which text degree of similarity is computable, is associated as a retrieval tag; an information selection unit for comparing a feature of each spoken content item extracted from each item of said speech information, with a feature of spoken content extracted from said input speech, to select information with which speech information similar to input speech is associated. The system further comprises an output unit for outputting information selected by said information selection unit, as information associated with input speech.

Methods and systems for improved license plate signature matching by similarity learning on synthetic images comprise generating a plurality of synthetic license plate images; applying one or more transformations to the synthetic license plate images to cause the synthetic license plate images to more closely resemble authentic license plate image captures; and providing the synthetic license plate images as inputs to a machine distance learning algorithm in which weighted similarity scores are calculated between signatures of analogous and non-analogous license plate images and one or more sets of signature weights are iteratively adjusted to increase the likelihood that comparing analogous license plate images results in high weighted signature similarity scores and comparing non-analogous license plate images results in low weighted signature similarity scores.

The embodiments of the invention disclose a face identification method and apparatus. First-type characteristics are obtained from pictures comprising human faces, then a distance weight coefficient between two faces is obtained according to the first-type characteristics, a weight similarity between the two faces is obtained by use of the distance between the two faces and the corresponding distance weight coefficient, and finally, according to the weight similarity between the two faces, clustering is performed on the pictures, the pictures belonging to the same person are clustered to form a cluster, and the pictures of the different persons are classified into different clusters. According to the face identification method, the similarity of five sense organs and facial forms of the two faces is taken into consideration, and measurement is also carried out on the similarity of the faces according to the first-type characteristics included in the pictures, such that the face similarity of the same person under different conditions is increased, and the similarity between the faces of different persons can be reduced. Therefore, the face identification method and apparatus can improve the face identification accuracy.

A speech recognition apparatus predicts, based on the occurrence cycle and duration time of impulse noise that occurs periodically, a segment in which impulse noise occurs, and executes speech recognition processing based on the feature components of the remaining frames excluding a feature component of a frame corresponding to the predicted segment, or the feature components extracted from frames created from sound data excluding a part corresponding to the predicted segment.

The embodiment of the invention relates to the field of data processing and in particular relates to a community partitioning method and device based on a characteristic matching network, which are used for partitioning communities. According to the embodiment of the invention, the community partitioning method comprises the following steps: determining K-bit Hash vectors corresponding to each piece of account information according to pre-set K Hash functions; sequentially dividing the Hash vector corresponding to each account information into m=K / h types of sub-Hash vectors; aiming at each type, dividing the account information with the same sub-Hash vector into the same group; calculating the similarity among all the account information in the same group; if the similarity among all the account information is more than a threshold value, establishing an interconnection edge among all the account information, so as to form the characteristic matching network; carrying out the community partitioning on the account information according to the characteristic matching network, so as to carry out the community partitioning according to the partitioned community and abnormal communities are found.

The invention discloses a fingerprint collection method and apparatus. The fingerprint collection method comprises the steps of determining damaged induction units in a plurality of induction units included in a fingerprint sensor; and when a fingerprint collection event is monitored, collecting fingerprint images by the multiple induction units, and setting the fingerprint image values collected by the damaged induction units to be random image values so as to form the fingerprint images collected by the fingerprint sensor. According to the technical scheme of the embodiments of the fingerprint collection method and apparatus, the similarity of the fingerprint image values collected by the damaged induction units in different time is reduced, so that the error probability in recognizing fingerprints is reduced.

A minimal spanning tree-based clustering genetic algorithm complex web community mining method belongs to the field of complex web community mining technology and is characterized in that the method comprises the following steps: carrying out computer initialization, carrying out population initialization, clustering population by a minimum spanning tree method, carrying out single-point crossover operation, mutation operation and selecting operation on individuals after population clustering, and carrying out iteration for T times so as to obtain the optimum community division of complex networks. Through minimal spanning tree-based clustering of populations and by crossover among populations, population diversity is maintained, and immaturity convergence is inhibited. By crossover operation of excellent individuals in species, probability of searching spaces with better solutions is increased. By selecting a neighbor-node for maximizing local modularity M1 as a variance value, search efficiency of the algorithm is improved.

Provided are an adjacent intersection arterial road coordinate control model and an optimization method thereof. The method includes the steps of building an adjacent intersection four-phase arterial road coordinate control model; performing individual coding, initializing data and setting parameters; performing population initialization; calculating values of the fitness of individuals in a population; performing minimum spanning tree clustering on the population; selecting individuals in the population to participate in genetic manipulation; performing genetic manipulation and mutation on the selected individuals; repeating iteration until obtaining the best individual. According to the adjacent intersection arterial road coordinate control model and the optimization method of there adjacent intersection arterial road coordinate control model, the minimum total delay of each vehicle team waiting in line is set as performance indicators, a more complete adjacent intersection arterial road coordinate control model is built by analyzing the signal states that front vehicles and rear vehicles of a vehicle team released in different phases at the upstream intersection encounter before reaching the downstream intersection and whether all vehicles at the intersection are released during release or not, an improved genetic algorithm is used as an optimization algorithm, and the length of the cycle, the arterial road timing, the signal timing and the phase difference are used as parameters for optimization solution.

The invention discloses a rain removing method for a single image, and the method comprises the steps: firstly decomposing a rain image into a high-frequency image and a low-frequency image based on bilateral filtering; secondly introducing the incoherence of a dictionary at a dictionary learning state so as to reduce the similarity between atoms with rain and atoms without rain, and constructinga new target function, thereby guaranteeing the separability of a rain dictionary and a no-rain dictionary during clustering. Moreover, the learning incoherent dictionary has the attributes similar tothe attributes of a tight frame, and can approach to an equal-angle tight frame. The method achieves the sparse expression of the high-frequency image through the rain dictionary and the no-rain dictionary, can separate a rain component and a no-rain component in the high-frequency image in a better way, achieves the superposing of the no-rain component of the high-frequency image with the low-frequency image, and achieves the removing of rain of the image. An experiment result indicates that the incoherent dictionary learned through the method is better in sparse expression performances, there are fewer the residual rain lines on the image after rain removal, the method maintains the edge details in a better way, and the visual effect is clearer and more natural.

The embodiment of the disclosure discloses a face identification method and apparatus. According to the face identification method, with cloths feature information of figures contained in a photo and / or photo shooting time information, weighting is carried out on a face distance of two figures. The more similar the cloths feature information of the two figures are, the shorter the face distance of the two figures will become, so that the similarity level of the two figures is increase; and the more different the cloths feature information of the two figures are, the longer the face distance of the two figures will become, thereby reducing the similarity level of the two figures. In conclusion, according to the method, similarity between face features of the five sense organs and faces of the two faces are taken into consideration and the influence on the face similarity of the photo feature information is also considered, thereby improving the accuracy of the face identification.

A device for removing haze from an image includes an image decomposer that decomposes a near-infrared (NIR) image to generate an NIR detail layer image and decomposes an RGB image to generate an RGB detail layer image and an RGB base layer image, a weight generator that generates a mixing weight value based on a similarity between high frequency (HF) components of the NIR image and the RGB image, a detail layer mixer that mixes the NIR detail layer image and the RGB detail layer image based on the mixing weight value to generate a mixed RGB detail layer image, a base layer dehazer that removes haze from the RGB base layer image to generate a compensated RGB base layer image, and an adder that adds the mixed RGB detail layer image and the compensated RGB base layer image to generate an output RGB image.

A method includes steps of: receiving data from a network by a web browser executed by a computer connected to the network, analyzing the data received for determining if the data received is transmitted by a network device, extracting keywords from the data transmitted by a network device, generating a query from the keywords extracted from the data received, sending the query through the network to a web interface of a support database providing support information about operating the network device, receiving from the network query results transmitted by the support database, and displaying the query results with the page transmitted by the network device.

In a sound processing apparatus, a matrix factorization unit acquires a non-negative first basis matrix including a plurality of basis vectors that represent spectra of sound components of a first sound source, and acquires an observation matrix that represents time series of a spectrum of a sound signal corresponding to a mixed sound of the first sound source and a second sound source different from the first sound source. The matrix factorization unit generates a first coefficient matrix, a second basis matrix and a second coefficient matrix from the observation matrix by non-negative matrix factorization using the first basis matrix. A sound generation unit generates either of a sound signal according to the first basis matrix and the first coefficient matrix or a sound signal according to the second basis matrix and the second coefficient matrix.

The invention relates to an OCR recognition method based on a deep learning model and a terminal, and belongs to the field of data processing. The method comprises the steps of segmenting a preset character segment image into a plurality of single-character images to obtain a single-character image set; sequentially identifying elements in the single-character image set by a preset first OCR deeplearning model to obtain a first feature vector set; wherein one single-character image corresponds to a first feature vector; according to a preset feature database, each first feature vector in thefirst feature vector set is converted into a corresponding single character, and a single character set is obtained; wherein a record in the feature database stores a single character and a feature vector corresponding to the single character; and arranging elements in the single character set to obtain a character string corresponding to the character segment image. The anti-interference capability of OCR character recognition is improved.

The invention relates to a method for controlling the automatic switching of the projector of a vehicle, in particular of a motor vehicle, from a first lighting mode to a second lighting mode, the method being implemented when the vehicle is driven at night and including the following steps: acquisition by a black and white sensor of images representing night-time road scenes, extraction from the images of the zones corresponding to light sources, and discrimination of the type of light sources in accordance with the distribution of the levels of light intensity in the zones.

The invention provides an improved multichannel spectral clustering algorithm-based cleaning robot map segmentation method. The method comprises the following steps that: parameters are inputted; a distance transformation algorithm is called to calculate distances between every two idle grids in a grid map, and a distance matrix is constructed; a Gaussian kernel function is adopted to construct a corresponding similarity matrix on the basis of the distance matrix, and a degree matrix is constructed according to the similarity matrix; a standardized Laplacian matrix is calculated on the basis of the similarity matrix and the degree matrix; an eigenmatrix is constructed according to eigenvectors corresponding to the first k largest eigenvalues of the Laplacian matrix; the eigenmatrix is standardized, so that an eigenmatrix which is represented by a symbol described in the descriptions of the invention can be obtained, with each line of the eigenmatrix which is represented by the symbol described in the descriptions of the invention adopted as one k-dimension sample, an algorithm is adopted to perform clustering; if vectors in the m-th line of the standardized eigenmatrix which is represented by the symbol described in the descriptions of the invention are allocated to an n-th cluster, an m-th idle grid is allocated to an n-th sub-region; and a segmented grid map is outputted. According to the method of the present invention, the influence of adjacent grids is fully considered, and the adaptability of the algorithm is improved.

The invention relates to the technical field of molecular biology detection, in particular to a TaqMan-MGB fluorescent quantitative polymerase chain reaction (PCR) detection specific primer (as shown in SEQ ID No. 1 and SEQ ID No. 2), a TaqMan-MGB fluorescent quantitative PCR detection probe as shown in SEQ ID No. 3) and a TaqMan-MGB fluorescent quantitative PCR detection method for seneca valley virus (SVV). The method comprises the steps of drawing a standard curve; extracting ribonucleic acid (RNA) of a sample virus; carrying out reverse transcription on the RNA of the sample virus; enabling the product of the reverse transcription to have a TaqMan-MGB fluorescent quantitative PCR, and reading the result. The primer provided by the invention has better specificity and sensitivity; the MGB probe provided by the invention is shorter and is beneficial to probe design, and the Tm value difference between a paired template and a non-paired template is improved, so that the experimental result is more stable and accurate; the method provided by the invention has the advantages of being simple and rapid, easy to operate, visual in results, high in sensitivity, good in stability, real-time quantitative, and the like, and shortens the reaction time.

The invention relates to the technical field of detection of pathogenic microorganisms, in particular to a detection marker, primer probe pair, kit and detection method for herpesvirus hominis I typeand II type. By taking reverse repeated sequences RS1 and RS2 in an HSV genome as a target area, the detecting sensitivity can be further improved as two copies are available in the genome on the onehand and cross reactions are unlikely to be caused as the inter-nucleic acid similarity is low on the other hand.