1844 results about "Coefficient matrix" patented technology

A method and apparatus for performing video compression and decompression using dynamic quantization and/or encoding. According to one aspect of the compression method, an image is recursively filtered into its constituent components each represented by a matrix of coefficients. Next, the level of correlation of a first of the matrices of coefficients is determined. Based on the level of correlation of this first coefficient matrix, one of a first quantization technique and a second quantization technique is selected. The first coefficient matrix is then quantized using the selected quantization technique.According to another aspect of the compression method, an image is digitally filtered into its constituent components, each represented by a matrix of coefficients. At least certain matrices are quantized to generated a first quantized matrix. The first quantized matrix is then recursively divided until the first quantized matrix or each of the resulting submatrices is sufficiently dense or sufficiently sparse. Each of the sufficiently dense matrices are encoded using a first technique, while each of the sufficiently sparse matrices are encoded a second technique.

A method and apparatus encode and decode video by determining whether to use discrete cosine transform (DCT) and DST for each of the horizontal and vertical transforms. During encoding, an intra-prediction is performed based on an intra-prediction mode determined for an M×N input image block to obtain an M×N intra-prediction residue matrix (E). Based on the intra-prediction mode, each of a horizontal transform and a vertical transform is performed using one of DCT and DST according to the intra-prediction mode. During decoding, the intra-prediction mode is determined from an incoming video bitstream. The M×N transformed coefficient matrix of the error residue is obtained from the video bitstream using an inverse quantizer. Based on the intra prediction mode, one of DCT and DST is performed for each of an inverse vertical transform and an inverse horizontal transform.

A method for decoding a multi-audio-object signal having audio signals of first and second types encoded therein, the multi-audio-object signal having a downmix signal and side information having level information of the audio signals of the first and second types in a first predetermined time/frequency resolution, the method including computing a prediction coefficient matrix C based on the level information; and up-mixing the downmix signal based on the prediction coefficients to obtain a first and/or a second up-mix audio signal approximating the audio signals of the first and second types, respectively, wherein up-mixing yields the first and/or second up-mix signals S₁ and S₂ from the downmix signal d according to a computation representable by

with “1” denoting—depending on the number of channels of d—a scalar, or an identity matrix, and D⁻¹ being a matrix uniquely determined by a downmix prescription according to which the audio signals of the first and second types are downmixed into the downmix signal, and which is also included by the side information, and H being a term independent from d.

A method and system using branching on hyperplanes and half-spaces computed from integral adjoint and/or kernel or dual lattice basis for solving mixed integer programs within specified tolerance. It comprises steps of: (1) preprocessing to ensure feasibility and linear objective; (2) computing adjoint and/or kernel lattice basis of the equality constraint coefficient matrix, or its transformed sub-matrix; (3) generating a generalized-branch-and-cut tree; (4) selecting a node and adding new constraints or approximating existing constraints; (5) processing a node to update lower and upper bounds, deleting nodes, or removing variables; (6 optional) computing an ellipsoidal approximation of continuous relaxation of (5); (7 optional) computing new lattice basis; (8) partitioning the set in (4) generating two or more nodes; (9) repeating (5-8) till termination. Such can be applied to problems in marketing management, data mining, financial portfolio determination, product design optimization, and other complex systems where optimization of system is desired.

In performing SVD-MIMO transmission, a set-up procedure is simplified while assuring a satisfactory decoding capability with a reduced number of antennas. A transmitter estimates channel information based on reference signals sent from a receiver, determines a transmit antenna weighting coefficient matrix based on the channel information, calculates a weight to be assigned to each of components of a multiplexed signal, and sends, to the receiver, training signals for respective signal components, the training signals being weighted by the calculated weights. On the other hand, the receiver determines a receive antenna weighting coefficient matrix based on the received training signals.

The invention is suitable for the technical field of displayers, and provides a method and device for removing liquid crystal displayer mura. According to the method and device for removing the liquid crystal displayer mura, brightness values of pixel points corresponding to input images are obtained, the brightness values of the pixel points corresponding to the input images are amended according to a brightness amendment coefficient matrix, the amended brightness values of the pixel points corresponding to the input images are obtained, then amended gray-scale values of the pixel points corresponding to the input images are obtained according to the amended brightness values of the pixel points corresponding to the input images, images corresponding to the amended gray-scale values are displayed on a liquid crystal screen, the liquid crystal displayer mura is effectively removed, the displaying effect of the liquid crystal screen is improved, and the viewing comfort senses of users are improved.

PCT No. PCT/JP96/03778 Sec. 371 Date Apr. 17, 1998 Sec. 102(e) Date Apr. 17, 1998 PCT Filed Dec. 25, 1996 PCT Pub. No. WO98/11456 PCT Pub. Date Mar. 19, 1998This invention provides a contaminant-detecting apparatus having high selectivity and high sensitivity against a contaminant. A product (1) is conveyed to a point where it is irradiated by x-rays from source (3). An x-ray detector (4), having a predetermined detection unit width in a direction perpendicularly intersecting the conveying direction, then detects the x-rays transmitted through the product. A storage unit (5) stores a two-dimensional distribution of x-ray intensity detected by the x-ray detector as a transmission image in units of pixels. An average calculation unit (7) performs a sum-or-product operation of a kernel, which is equal to or larger than 7x7 pixels, (9x9 or 11x11), and equal to or smaller than (a pixel count corresponding to +E,fra 1/2+EE the predetermined x-ray detection unit width)x(pixel count corresponding to +E,fra 1/2+EE the predetermined x-ray detection unit width), and includes a target pixel, in units of pixels of the transmission image stored in the storage unit by using a predetermined coefficient matrix, thereby calculating the weighted average over the kernel. A difference calculation unit (8) calculates the difference between the x-ray intensity of the target pixel of the transmission image stored in the storage unit and the weighted average over the kernel of the target pixel which is calculated by the difference calculation unit. A determination unit (9) compares the difference calculated by the difference calculation unit with predetermined criteria, thereby determining presence/absence of a contaminant in the product to be tested.

A thermostat may include a plurality of heat-generating components; a plurality of first temperature sensors, each of the plurality of first temperature sensors being disposed next to a corresponding one of the plurality of heat-generating components; a second temperature sensor that is disposed away from the plurality of heat-generating components; and a memory device storing a coefficient matrix. The thermostat may also include one or more processors that combine a plurality of inputs to calculate an ambient temperature for an enclosure in which the thermostat is installed, the plurality of inputs including readings from the plurality of first temperature sensors, readings from the second temperature sensor, and the coefficient matrix.

The invention provides a method based on linear programming for locating near-field targets and a system thereof, particularly a method for locating near-field targets on the basis of the compressivesensing theory. The method comprises the following steps: selecting a reference array element under the condition that the signal form is unknown, allowing the reference array element to work at the normal sampling frequency, and the other array elements to work at far below the Nyquist sampling frequency; taking the output signal of the reference array element as the reference target signal, to acquire the sample data of all the array elements; generating a time-delay table and a sparse basis array; generating a random mapping array and obtaining a coefficient array; obtaining a sparse vectorby linear programming solution; and acquiring the location information of the near-field target from the predetermined location-distance collection on the basis of the acquired estimation results ofthe sparse vector. According to the invention, the sensor does not need to work beyond the Nyquist sampling frequency, thereby greatly reducing the sampling rate, reducing the operating energy consumption of the sensor and improving the resource utilization rate of the system; and the method has no limits to the target bandwidth, so that the method is applicable to the target location of both narrowband and wideband, and the method is further applicable to non-Gaussian target location.

The invention dislcoses a single-phase earth fault line selection method based on transient state high-frequency component correlation analysis. The method comprises the steps that whether or not a single-phase earth fault happens is judged according to the magnitude of a busbar zero-mode voltage instantaneous value, db10 wavelets are adopted for performing five-layer wavelet packet decompositionon zero-mode current of two work frequency periods after each feeder line fault, after the minimum frequency band is removed, node wavelet packet deposition coeffiences are reconstructed and summeredto obtain a fault transient high-frequency capacitance-current component, and a correlation coeefient matrix M between every two feeder lines is obtained; Si is set as a correlation accumulation coefficient for the ith feeder line fault transient high-freuqency capacitance-current component, when rho ij is larger than rho set, Si is equal to 1, 1 is subtracted from Si, and a correlation accumulation coefficient matrix S can be obtained; if Si is larger than 0 constantly, the busbar breaks down; if Si of only one feeder line is smaller than 0, the circuit is a fault circuit; if Si is smaller than 0 constantly, the circuit with the maximum high-frequency-band wavelet packet energy sum is determined as the fault circuit. According to RTDS simulation experimental verification, the method has the good accuracy, adaptability and the like.

The invention discloses a multi-target distribution and flight path planning method for multiple rescue helicopters, which relates to the technical field of aviation emergency rescue. The method comprises the steps of evaluating aviation rescue task execution feasibility, conducting disaster area map image processing, obtaining a feasibility matrix through the evaluation of the rescue task execution feasibility, taking the feasibility matrix as a coefficient matrix for a solution matrix Xij, solving distribution results according to tasks of different quantities of the rescue helicopters and target points through a Hungary algorithm, obtaining a flight path solution space by extracting a safety zone topological structure by adopting an image processing method and morphological characteristics, obtaining the shortest path in the flight path solution space by a heuristic A* search algorithm, combining the results of multi-target distribution for multiple rescue helicopters, forming an emergency plan for rescue task execution of the helicopters, and calling a Google Earth flight simulation API (Application Program Interface) to demonstrate the emergency rescue plan and making preflight preparation, so as to improve the efficiency and safety of the aviation emergency rescue. The method is simple, the time complexity is low, the flight path planning process is visual, and the result is distinct.

The invention relates to an image quality assessment method, in particular to a non-reference image quality assessment method based on information entropy characters, and belongs to the field of image analyzing. The method comprises the first step of carrying out Contourlet conversion on a distorted image to obtain N*M conversion sub-bands, the second step of dividing each conversion sub-band and the unconverted original distorted image, the third step of calculating null domain information entropy and frequency domain information entropy on each block coefficient matrix, and the fourth step of screening the blocking characters and calculating a mean value to obtain the quality character value of each conversion sub-band. The method of a support vector machine and the method of non-reference image quality assessment are utilized for testing on a test set, and quality prediction and assessment are carried out through quality character vectors corresponding to a disaggregated model, an evaluation model and the test set all of which are obtained through training. The non-reference image quality assessment method has the advantages of being high in subjective consistency, small in time complexity and good in university, can be embedded into application systems related to image quality, and has very high application value.

The invention discloses a near infrared spectrum analyzing method based on the isolated component analysis and genetic neural network, which comprises the following steps for the acquired near infrared spectrum: firstly, effectively compressing spectrum data by using wavelet transform; secondly, extracting an independent component and a corresponding mixed coefficient matrix of a near infrared spectrum data matrix by using an isolated component analysis method; thirdly, building a three-layer BP neutral network, using the mixed coefficient matrix of a training sample as the input and correspondingly measured component concentration matrix as the output, and optimizing a neutral network structure by adopting a genetic algorithm, and obtaining a GA-BP neutral network by the training of the training sample; fourthly, predicting and analyzing the measured component concentration of the predicted set sample by using the GA-BA neutral network. The method enriches the chemical measurement method, widens the application range of the isolated component analysis and has favorable application prospect.

The invention discloses a fast decoupled flow calculation method for power systems, which comprises the following steps of: inputting original data and initializing voltage; forming an admittance matrix; forming correction equation coefficient matrixes B' and B'' and performing factor table decomposition; performing P-theta iteration, and correcting a voltage phase angle; performing Q-V iteration, and correcting voltage amplitude; judging whether the iteration is converged; and calculating node power and branch power. The method requires that the P-theta iteration and the Q-V iteration are all converged in the same iteration and the iteration process is finished, so that the algorithm frame is simpler, and the flow is clearer. The sparse matrix technology is not adopted, so the matrix elements are convenient to access and calculate, and the programming is simple; the correction equation coefficient matrixes are stored according to n order, number change of nodes is avoided, and the programming difficulty is reduced; and the calculation amount is reduced through reasonable logic judgment, the calculation speed is obviously improved and the requirement of scientific research can be completely met. The fast decoupled flow calculation method also can process power systems with a plurality of balance nodes.

In methods and systems that may facilitate detecting abnormal operation in a process plant, values of a plurality of process variables may be analyzed to determine whether any of a plurality of faults associated with the process plant exist. If one or more faults are detected, one or more indicators may be generated. Analyzing the values of the plurality of process variables may include utilizing a coefficient matrix. The coefficient matrix may be generated based on process variable data corresponding to the known occurrences of faults.

The invention discloses a multi-label text classification method and system. The method comprises the following steps: obtaining a training set comprising a text sequence and a label space, extractingglobal feature vectors of all words in the text sequence by adopting a long-short time memory network, and aggregating the obtained global feature vectors by adopting a convolutional neural network to obtain a semantic vector of each word in the text sequence; respectively calculating weight coefficients of each label in the note space and all words in the text sequence, constructing an attentionweight coefficient matrix, and processing the attention weight coefficient matrix to obtain an optimal weight coefficient matrix; respectively weighting the semantic vector of each word and the weight coefficient vector in the optimal weight coefficient matrix to obtain an attention vector of the tag; and performing normalization processing on the attention vectors of the tags to obtain the probability of each tag, and selecting several tags with maximum probabilities to classify the text.

The invention provides a hypergraph convolutional network model and a semi-supervised classification method thereof. The method comprises the following steps: 1, carrying out sparse coding on sample data features of a non-Euclidean structure to form a sample representation coefficient matrix; 2, constructing a hyperedge according to the similarity of the samples, calculating a hyperedge weight, and constructing a hyper-graph model; 3, by means of a hypergraph theory, defining convolution operation on a hypergraph, and constructing a hypergraph convolution network model; 4, defining a semi-supervised learning method on the hypergraph convolutional network, designing a loss function, and predicting category labels of all samples by using category information of a small number of calibrationsamples; 5, respectively making a semi-given label matrix for training, verifying and testing, setting network hyper-parameters, training a network model, and learning a convolution kernel and a regularity factor parameter of the network according to a random gradient descent algorithm; and 6, for given data, predicting an unknown sample category by using the trained model to realize semi-supervised classification.

The invention discloses a multiple time phase remote sensing image registration method based on an edge and spectral reflectivity curve. The multiple time phase remote sensing image registration method based on the edge and spectral reflectivity curve mainly solves the problems that stability is insufficient and accuracy is low when registration is conducted on two time phases of multiple spectral image sets with variation regions and whole luminance difference in the prior art. The multiple time phase remote sensing image registration method based on the edge and spectral reflectivity curve comprises the following steps: small wave high frequency coefficient matrixes are added and divided, the small wave high frequency coefficient matrixes are added simultaneously, the input two time phases of the multiple spectral image sets are preprocessed, and thus two time phases of strong edge images are acquired; registration parameters of edge alignment metric of the two time phases of the strong edge images are calculated to register the two time phases of image sets and acquire coarse registration image sets; the coarse registration image sets are converted into opposite object spectral reflectance image sets, small wave high frequency coefficients of time phases 1 of the object spectral reflectance image sets are reconstructed to acquire coarse registration edge images, and the coarse registration edge images conduct local matching on opposite object spectral reflectance images of the time phases 1 and time phases 2 to acquire a final registration image set. The multiple time phase remote sensing image registration method based on the edge and spectral reflectivity curve can be used for change detection, integrating and embedding.

A method and apparatus is provided for implementing quantization in encoding/decoding. Firstly, the quantization on the picture transformed coefficient matrix is performed using a quantization parameter model, that is, the coefficient matrix is modeled as a parameter model represented by only several parameters, in which the parameter model includes two sorts of parameters, frequency band distribution parameters and frequency band weighting parameters. Thus, a quantization matrix conforming to the human vision system can be obtained by controlling the model parameters and the subjective quality of the encoded picture can be easily controlled. Secondly, the method does not need to store or transfer the quantization matrix in the bitstream which help to improve the encoding efficiency. Finally, the quantization on picture transformed coefficients using the method is well adapted to the content information of the picture to be encoded. Therefore, the present invention can improve the subjective quality of the encoded picture.

The invention discloses a standard dot matrix light source and an image point correction method of a light-emitting diode (LED) display screen. The method comprises the following steps of: A, constructing the standard dot matrix light source, and calibrating the brightness and the chroma of the standard dot matrix light source; B, acquiring image data of the LED display screen, and the brightness and the chroma of the standard dot matrix light source; C, comparing the acquired brightness and chroma of the standard dot matrix light source with a stored value to obtain an error compensation matrix, and converting the image data of the LED display screen from a relative value into an absolute value; D, analyzing the converted image data of the LED display screen to generate a data correction coefficient matrix; and E, outputting the data correction coefficient matrix to an image point correction processor to perform image data correction processing so as to obtain corrected image data according to the data correction coefficient matrix, and finally outputting the corrected image data to an LED large screen display system. The standard dot matrix light source and the image point correction method can overcome adverse effects caused by working condition drifting of a camera to ensure that the LED display screen has better light-color consistence and white balance after performing the image point correction; and an additional light colorimeter is not needed, and a plurality of times of tests are not performed any more.

The invention discloses a face recognition method based on weighted diagnostic sparseness constraint nonnegative matrix decomposition, and mainly aims to solve the problem that the method in the prior art is not robust to an obscured face and is of low recognition rate. According to the technical scheme, the method comprises the following steps: (1) constructing a nonnegative weight matrix according to the obscured area of a test image; (2) introducing the weight matrix into a general KL divergence objective function, applying a sparseness constraint to a basis matrix, and applying intra-class and inter-class divergence constraints to a coefficient matrix to get a weighted diagnostic sparseness constraint nonnegative matrix decomposition objective function; (3) solving the objective function, and decomposition-training a data matrix to get a basis matrix and a coefficient matrix; (4) projecting a test data matrix on the basis matrix to get a corresponding low-dimensional representation set, and taking the low-dimensional representation set as final test data; and (5) using a nearest neighbor classifier to classify the test data by taking the coefficient matrix as training data, and outputting the result. By using the method, the effect of obscured face recognition is improved. The method can be used in identity recognition and information security.

The invention provides for the encoding of surround sound produced by any coincident microphone techniques with coincident-to-virtual microphone signal matrixing. An encoding scheme provides significantly lower computational demand, by deriving the spatial parameters and output downmixes from the coincident microphone array signals and the coincident-to-surround channel-coefficients matrix, instead of the multi-channel signals.

The invention relates to an image encryption method and an image decryption method with visual security and data security based on compressed sensing. The image encryption method comprises the steps of: firstly, utilizing an SHA 256 hash function to obtain a 256-bit hash value of a plaintext image as an image secret key, and calculating initial numerical values of one-dimensional skew tent chaotic mapping and zigzag scrambling; carrying out sparse processing on the plaintext image, and carrying out zigzag scrambling on a coefficient matrix; and then utilizing the one-dimensional skew tent chaotic mapping to generate a measurement matrix, measuring and quantifying a scrambling matrix to obtain a compressed and encrypted image, and embedding the image into a carrier image with visual significance to obtain a final ciphertext image with visual significance. The image encryption method realizes the visual security and data security of the plaintext image, has large secret key space, is highly sensitive to plaintext, has higher capacity of resisting brute-force attack, chosen-plaintext attack and known-plaintext attack, does not need an additional storage space, and can transmit and store the ciphertext image quickly and effectively.