146 results about "Linear representation" patented technology

The invention relates to a graph-based semi-supervised high-spectral remote sensing image classification method. The method comprises the following steps: extracting the features of an input image; randomly sampling M points from an unlabeled sample, constructing a set S with L marked points, constructing a set R with the rest of the points; calculating K adjacent points of the points in the sets S and R in the set S by use of a class probability distance; constructing two sparse matrixes WSS and WSR by a linear representation method; using label propagation to obtain a label function F<*><S>, and calculating the label prediction function F<*><R> of the sample points in the set R to determine the labels of all the pixel points of the input image. According to the method, the adjacent points of the sample points can be calculated by use of the class probability distance, and the accurate classification of high-spectral images can be achieved by utilizing semi-supervised conduction, thus the calculation complexity is greatly reduced; in addition, the problem that the graph-based semi-supervised learning algorithm can not be used for large-scale data processing is solved, and the calculation efficiency can be improved by at least 20-50 times within the per unit time when the method provided by the invention is used, and the visual effects of the classified result graphs are good.

The invention discloses a hyperspectral abnormal object detection method based on structure sparse representation and internal cluster filtering, aiming at addressing the technical problem of low object detection effciency of current hyperspectral abnormal object detection methods. The technical solution involves: after selecting an initial background pixel, using the dictionary learning method which is based on principal component analysis to study a background dictionary which obtains rebustness, in the course of sparse vector resolution and image reconstruction, introducing re-weighted laplacian prior to increase the solution precision of sparse vector, computing the errors betwen an original image and a reconstructed image to obtain a sparse representation error, using the internal cluster filtering to represent space spectrum characteristics of hyperspectral data, obtaining the internal cluster error by computing the error between a to-be-tested pixel and other pixel linear representation result, and finally combining the sparse representation error and the linear weighting of the internal cluster error and implementing precise extraction of an abnormal object. According to the invention, the method increases 10-15% of detection rate with the proviso of a constant false alarm rate compared with prior art.

A method of validating binary code transformation in one aspect includes analyzing original program and transform program. Control flow graphs are generated for both programs. The two graphs are traversed to create respective linear invariant representations. The linear representations are compared to identify incorrect transformations.

The invention discloses an improved method based on an extreme learning machine (ELM) and sparse representation classification. The method comprises the following steps of 1, randomly generating a hidden layer node parameter; 2, calculating a hidden layer node output matrix; 3, according to a size relation of L and N, using different formulas to calculate an output weight connecting a hidden layer node and an output neuron; 4, calculating an output vector of a query picture y; 5, determining a difference value of a maximum value of and a secondary maximum value os in an ELM output vector o, and if the difference value is greater than a set value, determining an index corresponding to the maximum value in the output vector, wherein the index is a type which the query picture belongs to; otherwise, entering into step6; step6, using a training sample corresponding to the k maximum values in the output vector o to construct a dictionary, using a coefficient reconstruction algorithm to calculate a linear representation coefficient of the picture y, calculating a residual error and determining the type which the query picture belongs to according to the type corresponding to the residual error. In the invention, a calculated amount is greatly reduced, a high recognition rate is realized and calculating complexity can be greatly reduced.

Coding techniques for a video image compression system involve improving an image quality of a sequence of two or more bi-directionally predicted intermediate frames, where each of the frames includes multiple pixels. One method involves determining a brightness value of at least one pixel of each bi-directionally predicted intermediate frame in the sequence as an equal average of brightness values of pixels in non-bidirectionally predicted frames bracketing the sequence of bi-directionally predicted intermediate frames. The brightness values of the pixels in at least one of the non-bidirectionally predicted frames is converted from a non-linear representation.

The invention belongs to the technical field of artificial intelligence, and the domain adaptive method of the invention also belongs to an important branch of the migration learning field, and discloses a multi-source domain adaptive face recognition method. The paper presents a method of learning the common subspace of multiple source domains and a single target domain, which is applied to facerecognition problem. It solves the problem that the classifier which learns from the source domain has poor recognition effect in the target domain when the data of the source domain and the target domain are distributed differently and the data of the target domain does not have or has a small number of labels. By learning samples from multiple source domains and single target domains, a common subspace is obtained. By transforming the source domain into the linear representation of the target domain in the common subspace, the target source domain data is casted into the high dimensional space to construct the linear optimal hyperplane in the high dimensional space, and a kernel classifier is learned. Then the test samples are identified and classified, and the beneficial results are obtained.

The invention discloses a face recognition method based on kernel nearest subspace, mainly solving the problem that the non-linear characteristics of the data can not be subjected to linear expression in the existing methods. The method comprises the following steps: (1) mapping the training sample matrixes and the testing samples to the non-linear characteristic space by Mercer kernel experience, then carrying out dimension reduction and normalization on the mapped samples and then extracting each class of training samples undergoing dimension reduction; (2) solving the sample reconstruction coefficient between the normalized testing samples and each class of training sample matrixes and reconstructing the original testing samples; and (3) obtaining the residual errors between various classes of reconstructed samples and the original testing samples and taking the class of the subscript corresponding to the minimum in the residual errors as the class of the testing samples. The method improves the precision in face recognition application, simultaneously expands the application range to the low-dimensional samples so as to further have universality and can be used for supervision and protection of public security, information security and financial security.

The invention discloses a method for identifying second level human face with weighted collaborative representation and linear representation classification combined. The method combines CRC and LRC. The method includes the following steps: firstly based on PCA, reducing dimensions of all the image samples so as to reduce computing complexity; taking different contributions of sample local similarity prior information on identifying classifications into consideration, constructing a weighted matrix and embedding to the CRC, proposing the weighted CRC, and then based on the weighted CRC, in accordance with reconstruction residual error ordering, reserving plural types of training samples with a greater similarity for LRC so as to realize secondary classification identification. The method reduces classification objects, which makes identification more accurate, and substantially reduces time for identification.

The present invention is a method for identifying unknown parameters in a system having a set of governing equations describing its behavior that cannot be put into regression form with the unknown parameters linearly represented. In this method, the vector of unknown parameters is segmented into a plurality of groups where each individual group of unknown parameters may be isolated linearly by manipulation of said equations. Multiple concurrent and independent recursive least squares identification of each said group run, treating other unknown parameters appearing in their regression equation as if they were known perfectly, with said values provided by recursive least squares estimation from the other groups, thereby enabling the use of fast, compact, efficient linear algorithms to solve problems that would otherwise require nonlinear solution approaches. This invention is presented with application to identification of mass and thruster properties for a thruster-controlled spacecraft.

The scheduling of extraction of a resource within a mine is disclosed which comprises obtaining drill hole grade data relating to the resource and creating a plurality of different block models, with each block in each model having a resource grade and each of the block models honoring the drill hole grade data. An extraction schedule is determined based on a cut-off grade policy to increase expected net present value of the mine. The cut-off grade policy can be a variable cut-off grade and the scheduling using the variable cut-off grade is converted from a non-linear representation to a linear representation and a mixed integer programming formulation is used for simultaneous optimization of the extraction schedule and the cut-off grade.

Methods, apparatus, and systems may operate to more efficiently utilize data stored in an array of storage blocks organized as rows and columns of contiguous blocks, where non-linearity is present in the data. Activities may include organizing data to discard useless elements from storage blocks when transferring the data to a memory buffer, and perhaps compressing the data for increased memory density utilization. Additional activities may include reconstructing data stored in the memory buffer and using an image distortion formula to display a linear representation of the non-linear data.

The invention provides a movement data compression method based on sparse representation. The method comprises the following steps of: firstly, analyzing input movement data to generate a sparse representation dictionary; then, carrying out sparse linear representation on each frame of the movement data based on the sparse representation dictionary; and finally, utilizing a K-SVD algorithm to carry out iterative optimization on the dictionary and the sparse representation. According to the method disclosed by the invention, under the precondition of controlling a reestablishing error into a reasonable range, the higher compression ratio can be achieved and the movement data compression method is particularly suitable for compressing shorter movement data.

A method of providing information relating to a path being travelled using a navigation apparatus, involves displaying a path bar including a linear representation of at least a portion of a path to be travelled using a navigation apparatus. The linear representation of the portion of the path is scrollable in the display window by a user to reveal a linear representation of another portion of the path. The method may involve automatically increasing a scale of the linear representation of the path displayed as a given location represented along the path is approached, wherein the location is a location of a safety camera or event affecting traffic flow along the path.

A method for displaying an anatomical image of a coronary artery on a graphical user interface with information acquired from a plurality of intravascular image frames. The method may include detecting qualitative information from the plurality of intravascular image frames, creating one or more indicator(s) from the qualitative information detected, and determining a spatial relationship between the anatomical image and a plurality of acquisition locations of the plurality of intravascular image frames and generating its linear representation. The method also includes displaying the anatomical image of the coronary artery with the linear representation overlaid thereon on a display device and overlaying the one or more indicator(s) representing at least one type of qualitative information on the anatomical image along the linear representation.

A method, system, and computer programs for improving the image quality of one or more bi-directionally predicted intermediate frames in a video image compression system, where each frame comprises a plurality of pixels. In one aspect, the invention includes determining the value of each pixel of each bi-directionally predicted intermediate frame as a weighted proportion of corresponding pixel values in non-bidirectionally predicted frames bracketing the sequence of bi-directionally predicted intermediate frames. In one embodiment, the weighted proportion is a function of the distance between the bracketing non-bidirectionally predicted frames. In another embodiment, the weighted proportion is a blended function of the distance between the bracketing non-bidirectionally predicted frames and an equal average of the bracketing non-bidirectionally predicted frames. In another aspect of the invention, interpolation of pixel values is performed on representations in a linear space, or in other optimized non-linear spaces differing from an original non-linear representation.

What is disclosed is a novel system and method for converting a set of L*a*b* values to CMYK color space in which all CMYK formulations that produce each L*a*b* value are found and enumerated. The CMYK formulations are found through a search algorithm starting with the lightest L*a*b* values then visiting neighboring L*a*b* values until the entire L*a*b* color set has been processed. The CMYK space is tessellated into a set of pentahedrons, and for each L*a*b* value, an enclosing pentahedron is found and the CMYK values corresponding to where the locus of this point penetrates each surface is recorded. Adjacent pentahedrons are then visited and this process continues until the gamut boundary is reached. The result is a piecewise linear representation of the CMYK locus containing all values that will give the target L*a*b* value. The present method provides a flexible and powerful approach for solving color management problems.

The invention discloses a face recognition method based on non-convex low-rank decomposition and superposition linear sparse representation. The method comprises the following steps: 1, according to alow-rank matrix decomposition theory, replacing a nuclear norm with a gamma norm for low-rank matrix decomposition, and introducing a structure incoherent discrimination item to form discriminative non-convex low-rank decomposition; 2, resolving the discriminative non-convex low-rank decomposition, and decomposing the face sample matrix into a low-rank matrix and a sparse matrix; 3, decomposing the low-rank matrix into prototype dictionaries and variation dictionaries through superposition linear representation, and then using the two dictionaries as dictionaries for testing through linear weighting combination; and 4, solving a sparse coefficient of l1 norm by using a homotopy method by using a sparse representation algorithm, carrying out classified recognition on the face images by reconstructing a minimum sparse residual model, and classifying the face samples to be tested into a class with a minimum error, thereby realizing face recognition. According to the method, good robustness and high efficiency can be maintained under the conditions of shielding and noise pollution.