1035 results about "Column vector" patented technology

PCT No. PCT / EP96 / 01272 Sec. 371 Date Nov. 21, 1997 Sec. 102(e) Date Nov. 21, 1997 PCT Filed Mar. 22, 1996 PCT Pub. No. WO96 / 29679 PCT Pub. Date Sep. 26, 1996The present invention concerns improved motion estimation in signal records. A method for estimating motion between one reference image and each frame in a sequence of frames, each frame consisting of a plurality of samples of an input signal comprises the steps of: transforming the estimated motion fields into a motion matrix, wherein each row corresponds to one frame, and each row contains each component of motion vector for each element of the reference image; performing a Principal Component Analysis of the motion matrix, thereby obtaining a motion score matrix consisting of a plurality of column vectors called motion score vectors and a motion loading matrix consisting of a plurality of row vectors called motion loading vectors, such that each motion score vector corresponds to one element for each frame, such that each element of each motion loading vector corresponds to one element of the reference image, such that one column of said motion score matrix and one motion loading vector together constitute a factor, and such that the number of factors is lower than or equal to the number of said frames; wherein the results from the Principal Component Analysis on the motion matrix are used to influence further estimation of motion from the reference image to one or more of the frames.

The invention discloses a convolutional neural network face recognition method based on multi-scale pooling. The method uses the multi-scale pooling-based convolutional neural network for extracting face image features to realize face recognition. During the convolutional neural network building process, a method with alternative convolution and maximum value sampling is adopted for carrying out deep extraction on the features, a multi-scale pooling strategy and a step are adopted for each convolution layer and are together inputted to a fully-connected layer, and thus, characteristic column vectors with multiple scales and a fixed size are provided. Cutting or size adjusting does not need to be carried out on the inputted face image, and images of different sizes can use the same network for training and recognition. According to the convolutional neural network based on the multi-scale pooling, the problem that the size of the input image can be not fixed is solved, the network can extract multi-scale face features, the network performance can be greatly enhanced, and wide application of the multi-scale pooling-based convolutional neural network to face recognition can be promoted.

A base station includes a storage unit to store a codebook, wherein the codebook includes a plurality of matrices; a control unit to perform beamforming using the codebook to generate a signal; and a multi-antenna to transmit the signal. For all of the plurality of matrices, all column vectors of a same matrix of the plurality of matrices are orthogonal to each other. For all columns of the plurality of matrices, all column vectors of a same column of all of the plurality of matrices satisfy a Grassmannian line packing (GLP) criterion.

The invention discloses a method for multi-input multi-output transmission for precoding. The method comprises the following steps: a fundamental matrix is selected, a column vector is selected from the fundamental matrix according to the dimension of a precoding matrix under each rank, the precoding matrixes are combined into codebooks, and the codebooks are respectively stored into the transmitting end and the receiving end of a multi-input multi-output transmission system; the receiving end selects precoding matrixes from the codebooks, and feeds back the indexes thereof to the transmitting end; the transmitting end uses the precoding matrixes corresponding to the indexes when the next transmission is performed, performs linear transformation in spatial domain to signals to be transmitted and then transmits the signals. The invention also discloses a method for multi-input multi-output codebook coding for precoding. The method enables the system to meet the feedback expenses and obtains good performance gain.

The invention discloses a clustering-based blind source separation method for synchronous orthogonal frequency hopping signals. The method comprises the following steps of: acquiring M sampled paths of discrete time-domain mixed signals; obtaining M time-frequency domain matrixes of the mixed signals; preprocessing the time-frequency domain matrixes of the frequency hopping mixed signals; estimating frequency hopping moments, normalized mixed matrix column vectors and frequency hopping frequency; estimating time-frequency domain frequency hopping source signals by utilizing the estimated normalized mixed matrix column vectors; splicing the time-frequency domain frequency hopping source signals between different frequency hopping points; and recovering time-domain source signals according to time-frequency domain estimate values of the source signals. According to the method, the frequency hopping source signals are estimated only according to the received mixed signals of a plurality of frequency hopping signals under the condition of unknown channel information, and the frequency hopping signals can be subjected to blind estimation under the condition that the number of receiving antennae is smaller than that of the source signals; short-time Fourier transform is utilized, so that the method is low in computation amount; and the frequency hopping signals are subjected to blind separation, and meanwhile, a part of parameters can also be estimated, so that the method is high in practicability.

The invention discloses a compressed sensing-oriented block-sparse signal reconfiguring method, and particularly relates to a block-sparse signal reconfiguring algorithm, which aims to solve the problems that the optimization complexity of a mixed l2 / l1 optimization algorithm in the conventional block-sparse signal reconfiguring algorithm is relatively higher and that overmatching phenomenon is easily caused by a block-sparse matching pursuit algorithm or a block-sparse orthogonal matching pursuit algorithm. The method of the invention comprises the following steps of: correcting labels, in ameasurement matrix, of column vectors of a recovery matrix calculated in the iteration operation of the (l-1)th time by performing the iteration of the lth time, and for a block-sparse signal x with the block sparsity of K, reconfiguring the block-sparse signal x by performing the iteration for not more than K times. The method is applied to the reconfiguration of the block-sparse signal, particularly to the reconfiguration of a binary block-sparse signal.

A base station includes a storage unit to store a codebook, wherein the codebook includes a plurality of matrices; a control unit to perform beamforming using the codebook to generate a signal; and a multi-antenna to transmit the signal. For all of the plurality of matrices, all column vectors of a same matrix of the plurality of matrices are orthogonal to each other. For all columns of the plurality of matrices, all column vectors of a same column of all of the plurality of matrices satisfy a Grassmannian line packing (GLP) criterion.

The invention relates to a method for identifying facial expressions from a human face image sequence, belonging to the technical field of analyzing and identifying human facial expressions. The method of the invention comprises the following steps of: firstly, adopting a method for tracing feature points, sequentially extracting the displacement amount of the normalized facial key point and the length of the special geometrical characteristic for each frame image of the expression image sequence, and combining the data to form a characteristic column vector; secondly, sequentially arranging all characteristic column vectors of the sequence to form a characteristic matrix, wherein each characteristic matrix represents a facial expression image sequence; finally, comparing the similarities among the characteristic matrixes by using a canonical correlation analysis method, thereby determining the human face images to be identified into one of the basic expressions of happiness, sadness, fear, hate, surprise and anger. In the invention, the canonical correlation analysis method is successfully applied to identifying the human facial expressions, the dynamic information in the expression generation course is utilized effectively and the higher recognition rate and the shorter CPU computation time are acquired.

The invention belongs to the technical field of digital image processing and mode recognition, in particular to a face recognition method based on blocking consistency LBP (Local Binary Pattern) and sparse coding. The face recognition method comprises the steps of: firstly, segmenting a face image into 16 subdomains which are same in size according to a mode of 4*4, calculating a consistency LBP histogram with one pixel radius and 8 neighbors, connecting LBP histograms of the 16 subdomains into a column vector to be used as a characteristic vector of a face image; and representing images to be tested into a most sparse linear combination on a training set, and recognizing the face image. Compared with the traditional characteristic extraction and clustering algorithm. According to the invention, structure information of a face can be well extracted, and under the condition of a single training sample and shielding, higher recognition rate and robustness are shown.

A method and apparatus for transmitting / receiving feedback information in a multi-user multi-antenna system and a system supporting the same are provided, in which all possible combinations are created using column vectors included in a precoding codebook, column vectors are correlated in each of the combinations, column vectors comprising correlations exceeding a reference threshold are designated as similar vectors, at least two similar vector sets are formed with the similar vectors, and feedback information is generated based on the at least two similar vector sets and transmitted.

The invention discloses a machine learning-based wireless sensing motion identification method which comprises the following steps: a step of data collection, a step of data denoising operation, a step of feature extraction and a step of SVM model training and identifying operation. During data collection operation, absolute value of a group of CSI data collected on each sampling point is obtained and is read into a 30*Nr*Nt matrix form. A PCA mode is mainly adopted for the data denoising operation. Feature extraction operation can be conducted based on discrete wavelet transformation. To make SVM model training convenient, training samples are subjected to Kmeans clustering operation via the machine learning-based wireless sensing motion identification method, n clustering centers can be used as word bags, and voting operation is performed based on feature vectors and best matching items of all the word bags; when the matrix form feature vectors are converted into column vectors, SVM model training can be realized conveniently. The machine learning-based wireless sensing motion identification method is a human body behavior identification method which is high in identification accuracy and high in robustness for environment change.

A pre-coding method and system and a method for constructing pre-coding codebook are disclosed, including: a transmitting end and a receiving end storing information of a pre-coding codebook together, the receiving end selecting a codeword from the pre-coding codebook according to an estimated channel matrix and feeding back a serial number of the codeword to the transmitting end, and the transmitting end finding out the codeword according to the serial number and pre-coding, by using the codeword, a symbol block sent to the receiving end; wherein, at least eight codeword vectors or column vectors in at least eight codeword matrixes in the pre-coding codebook are obtained based on 8-dimensional vectors selected from a 8-dimensional vector set, and this 8-dimensional vector set is obtained by calculating from the part or all of 4-dimensional vectors in u1, u2, u3, u4, u5, u6, u7, u8, s, m, n.

The invention discloses a processor and a method for executing matrix multiplication on the processor. The processor comprises a data bus and an array processor composed of k processing units. The data bus is used for reading n columns of row vectors in sequence from M*N multiplicand matrixes, inputting the row vectors to the processing units in the array processor, reading n*k sub-matrixes from N*K multiplier matrixes, inputting column vectors of the sub-matrixes into the corresponding processing units in the array processor, and outputting results obtained through multiplication executed by the processing units. Each processing unit in the array processor is used for executing parallel vector multiplication on the input row vectors and the column vectors, and each processing unit comprises a Wallace tree multiplying unit composed of n multiplying units and n-1 adders. The implementation mode improves processing efficiency of matrix multiplication.

Channel state information in a closed-loop, multiple-input, multiple-output wireless networks is fed back from each mobile station to a base station by first determining a transmit covariance matrix R, and applying a singular value decomposition (SVD) R=UΣVH, where U, V are left and right singular vector matrices, Σ is a diagonal matrix with singular values. The matrix V includes column vectors V. A beamforming vector vmax=[1 exp(jΦ)exp(j2Φ) . . . exp(jΦ)] / √{square root over (N)}] is approximated by the column vector V having a maximum magnitude, where Φ is a real number. Then, only the angle Φ is fed back using a phase modulation mapping of the components exp(jΦ) onto the associated subcarrier.

An apparatus and method for Orthogonal Cover Code (OCC) generation, and OCC mapping, includes generating the first group of OCC sequences, which is used to generate the first group of OCC sequences; generating the second group of OCC sequences, which is used to mirror the columns of the first group of OCC sequences to generate the second group of OCC sequences; generating the third group of OCC sequences, which is used to perform cyclic shift on the column vectors of the first group of OCC sequences to generate the third group of OCC sequences; and generating the fourth group of OCC sequences, which is used to mirror the columns of the third group of OCC sequences to generate the fourth group of OCC sequences.

A secret sharing apparatus according to the present invention is based on a (k,n)-threshold scheme with a threshold of at least 4. The secret sharing apparatus generates a generator matrix (G) of GF(2) in which any k of n column vectors are at a full rank, divides secret information into n−1 pieces to generate divided secret data (K(1), . . . , K(n−1)), generates random data (U(0,1), . . . , U(k−2,n−1)), calculates the product of matrixes of the divided secret data, the random data, and the generator matrix (G), assigns the j×(n−1)+ith column of the calculation result to sharing partial data (D(j,i)) to calculate sharing partial data (D(j,1)), generates header information (H(j)), and individually distributes n pieces of sharing information (D(0), . . . , D(n−1)) made up of the header information (H(j)) and sharing partial data (D(j,i)) to n storage apparatuses.

The invention discloses a face detection method based on a picture geometric structure, comprising a training process of face models and a detecting process of face images. The method comprises the following: the step of the training process of face models is divided into a step of training sample normalization, a step of feature extraction, a partitioning step of dividing a sample by adopting a block with a proper size, a step of drawing all differential values acquired by operations into a feature column vector to submit to a classifier for learning, a learning process of a fall type support vector machine and a classification against a sample picture in each window by using a cascade classifier; and the step of the detection of face images is to mark a detected face. The face detection method solves the problem generally existed in the field that the prior art aims at the local structure of a picture but can not completely and accurately express global information of the picture, and can detect a face quickly and accurately.

The invention relates to a low-sidelobe emission directional diagram design method improving DOA estimated performance of a MIMO radar. The method comprises the following steps: setting an interesting area; ensuring column vectors of a beam space weighting matrix to satisfy pairing characteristics; ensuring signals at a receiving end to satisfy rotation invariance; constructing an optimized model of a beam space weighting matrix with directional diagram coupling characteristics, rotation invariance of the signals and equality of various array element emission powers as constraint conditions; introducing an auxiliary variable and relaxing a rank-1 constraint to be a positive semidefinite constraint through utilization of the semidefinite relaxing technology; and through utilization of an interior point method, obtaining an optimal solution of the relaxing problem; through a Gaussian randomization method, the beam space weighting matrix is solved; and DOA estimation is carried out on a target at the receiving end through the ESPRIT algorithm. The MIMO radar waveform design technology which is good in performance in the invention can raise the signal to noise ratio at the receiving end, provide an important theoretical basis and a specific realization method for raising the precision of MIMO radar angle estimation.

Disclosed is a wireless communication system which carries out spatial multiplexing communication forming multiple spatial channels between a pair of a transmitter with multiple antennas and a receiver with multiple antennas, having the transmitter and the receiver. In the system, the multiple spatial channels are divided into one or more groups, the transmitter transmits reference signals corresponding to spatial channels in succession in a time-division manner for each group, and the receiver acquires a channel matrix whose column vectors are transfer functions of the spatial channels obtained from the reference signals for each group and spatially demultiplexes a receive signal using a receive weight obtained based on the channel matrix.

The invention relates to an active power distribution network transient state real-time simulation system designing method based on an FPGA. The active power distribution network transient state real-time simulation system designing method can be implemented in an off-line simulation environment and an on-line simulation environment in a dividing mode, the off-line simulation environment is responsible for calculating the total clock period number-n<total> and simulation using time-t<total> calculated in each time step, the simulation step size deltat is set according to the real-time simulation using time-t<total>, relevant parameters calculated according to the simulation step size deltat and the read basic parameter information are uploaded to the on-line simulation environment based on the FPGA. Real-time simulation calculation is completed in the on-line simulation environment, the simulation state is controlled through a finite-state machine, each time step includes the steps of historical item current source column vector calculation of all elements, total historical item current source column vector formation of the elements and linear equation system solving and updating of the elements, and the historical item current source column vector calculation step of the elements and the updating step are completely independent and can be processed in a concurrent mode. According to the active power distribution network transient state real-time simulation system designing method based on the FPGA, the real-time performance of the transient state simulation process of the whole system is guaranteed, and the good feasibility and the good applicability are achieved.

The invention discloses a pre-coding method, a pre-coding system and a construction method of a pre-coding code book. A transmitting terminal and a receiving terminal are used for storing the information of the pre-coding code book together; a code word is selected from the pre-coding code book by the receiving terminal according to an estimated channel matrix; the sequence number of the code word is fed back to the transmitting terminal; and symbol blocks transmitted to the receiving terminal are pre-coded through the code word by the transmitting terminal according to the code word found by the sequence number, wherein column vectors in at least eight code word vectors or at least eight code word matrixes in the pre-coding code book are obtained on the basis of an eight-dimensional vector selected from an eight-dimensional vector set, and the eight-dimensional vector set is obtained by computing partial or all four-dimensional vectors of u1, u2, u3, u4, u5, u6, u7, u8, s, m and n. The pre-coding method is used for the pre-coding of eight-antenna MIMO systems when the channel matrix is a singular matrix, and can enhance the pre-coding property. The four-dimensional vectors used for computing the code word can be stored only for saving storage space.

Pulsed or mutual interference in a pulse radar system is ameliorated by a process that includes a pulse-Doppler filter, some frequency bins of which correlate with clutter. Pulsed interference is identified, and the corresponding column vectors of a pulse-Doppler-filter-equivalent matrix are identified. The row values of the corresponding interference-affected column vectors that correlate with clutter are nulled. The vectors are orthogonalized, and then converted to measurement and correction signal vectors for application to the pulse-Doppler received signals. Vector dot products of the measurement signal vectors with the received signals are calculated, to produce pulsed interference measurements that are nominally free of clutter. The pulsed interference measurements are combined with the correction signals which are then subtracted from the received signals to produce signals that are nominally free of pulsed interference. In one embodiment, the pulsed interference-free received signals are applied to a pulse-Doppler filter to suppress the clutter signals.

The invention relates to a device and a method for generating column vector beams. TEM00 model beams outputted by a laser is changed into linearly polarized light through a polarizer after the beams are reshaped and expanded by a beam expander and a diaphragm; the beams enter a spatial light modulator or a complex filter, and form an array through an analyzer; the beams pass through a 4f system consisting of a pair of Fourier transform lenses and aperture diaphragms; the beams are added into a phase retarder for phase modulation so as to generate TEM10 (or TEM01) model beams, and are synthesized to the column vector beams by using an interference system. The device and the method have the advantages of low requirement on incident beam, simple device, easy operation, high polarization degree of the output column vector beams and the like.

The invention relates to a method for simply and easily calibrating an industrial robot. The method comprises the following steps of: establishing an industrial robot kinematics model, and solving a robot kinematics pose transfer matrix; extracting a robot end position component, and calculating an error transfer matrix; operating a robot end clamp to contact four side edges of a standard cube, reading a code disc number of each joint of the robot, and calculating a position coordinate; changing the pose of the standard cube, reading the code disc number of each joint of the robot again, calculating the position coordinate, and repeating the step for many times; and listing position coordinate matrix equations with an identification parameter, and arranging the position coordinate matrix equations in a manner of Ax=b, wherein A is of a matrix type, x is a column vector with the identification parameter, and b is the column vector, and if A is not strange, a parameter to be identified is solved, and the original parameter is corrected by using the identification result. The method disclosed by the invention has the advantages of convenience, practicability, lower cost, more simpleness on operations, capability of reaching a higher precision, and unnecessity of buying high-precision equipment, such as a high-precision laser instrument and a vidicon.

Provided is a radio communication device which can reduce ISI caused by destruction of an orthogonal DFT matrix even when an SC-FDMA signal is divided into a plurality of clusters and the clusters are respectively mapped to discontinuous frequency bands. The radio communication device includes a DFT unit (110), a division unit (111), and a mapping unit (112). The DFT unit (110) uses the DFT matrix to execute a DFT process on a symbol sequence in a time region to generate a signal (SC-FDMA signal) of the frequency region. The division unit (111) generates a plurality of clusters by dividing the SC-FDMA signal with a partially orthogonal bandwidth corresponding to the vector length of some of the column vectors constituting the DFT matrix used in the DFT unit (110) and orthogonally intersecting at least partially. The mapping unit (112) maps the clusters to discontinuous frequency bands.

The invention discloses a binary chop tracking method of shortest paths in a network map, which is characterized in that the method comprises the following steps: extracting the step number of the shortest paths from a source node to a gathered node according to a shortest path step number matrix; reducing and separating the step number by half into the front step number and the back step number; searching nodes with the step number equal to the front step number in a source node row vector of the shortest path step number matrix, using the nodes as a front step number node set, searching nodes with the step number equal to the back step number in a gathered node column vector of the shortest path step number matrix, using the nodes as a back step number node set; taking an intersection set of the front step number node set and the back step number node set as middle nodes; using each node in the intersection as the reference; using the method for solving the middle node of the source node and the node, and the middle node of the node and the gathered node until the shortest path step number between the solved nodes is 1; and sequentially inserting the middle nodes into corresponding positions to obtain all shortest paths from the source node to the gathered node. The method of the invention is used for finding all shortest paths between two points, and the invention can provide various schemes for the preferential path selection of the network map abstracted from actual data.

The invention discloses a multi-source image fusion method based on the synchronous orthogonal matching pursuit algorithm. The multi-source image fusion method comprises the following steps: sampling a source image pixel by pixel in an overlapping manner into image blocks of the same size by a sliding window of the fixed size and expanding each image block by columns into column vectors; obtaining the sparse representation coefficient corresponding to each vector on the over-complete dictionary by the synchronous orthogonal matching pursuit algorithm; fusing the corresponding coefficient by the maximum absolute value method; inverse-transforming the fused sparse representation coefficient into the fusion result vector of corresponding to the vectors according to the over-complete dictionary; and restoring all the fusion result vectors to image blocks and re-constructing to obtain the fused image. The invention fully considers the intrinsic characteristics of the image sparsity and the method using sparse representation can more effectively present the useful information of each source image and achieve better fusion effect, therefore, the invention is of great significance and practical value to the post-processing and image display of various application systems.

The invention provides a time-frequency transform based method for identifying parameters of a synchronous generator. The method includes steps of S1. carrying out Laplace transform on a differential equation of a d axle and a q axle of a generator in time domain in a generator model to obtain an equation of the d axle and the q axle of the generator in frequency domain, and carrying out inverse Laplace transform to obtain an integral equation of the d axle and the q axle of the generator in the time domain; S. replacing state variable of the integral equation of the d axle and the q axle of the generator in the time domain; S3. dispersing the integral equation of the d axle and the q axle of the generator in the time domain into that measurable column vectors or integration of the measurable column vectors are equal to a matrix multiplying by column vectors of coefficients of the integration equation, and solving to obtain the column vectors of the coefficients of the integration equation; and S4. solving to obtain parameters of the generator model according to the relation between the column vectors of the coefficients of the integration equation and parameters of the generator. A differential equation of a generator is converted to an integration equation according to the time-frequency transform based method for identifying the parameters of the synchronous generator, stability of calculating a value is facilitated, and the problem of an unstable identification result is solved.

First, the method defines pilot frequency sequence (PFS) composed of column vectors of orthogonal matrix, which is Fourier matrix. Then, the method scrambles PFS, and the scrambling sequence ck satisfies |ck|=1, and scrambled PFS does not influence on requirement of orthogonality. The invention obtains extensive uses in areas of wireless broadband communication and mobile communication, especially in fourth generation mobile communication, multicarrier, and OFDM transmission system. The method is mainly in use for designing sending PFS, and keeping the sending PFS as a signal in normal module in order to raise accuracy for estimating channel of receiver so as to raise performance of receiver.

A method and apparatus for transmitting / receiving feedback information in a multi-user multi-antenna system and a system supporting the same are provided, in which all possible combinations are created using column vectors included in a precoding codebook, column vectors are correlated in each of the combinations, column vectors comprising correlations exceeding a reference threshold are designated as similar vectors, at least two similar vector sets are formed with the similar vectors, and feedback information is generated based on the at least two similar vector sets and transmitted.