69 results about "Orthonormal basis" patented technology

A video compression framework based on parametric object and background compression is proposed. At the encoder, an embodiment detects objects and segments frames into regions corresponding to the foreground object and the background. The object and the background are individually encoded using separate parametric coding techniques. While the object is encoded using the projection of coefficients to the orthonormal basis of the learnt subspace (used for appearance based object tracking), the background is characterized using an auto-regressive (AR) process model. An advantage of the proposed schemes is that the decoder structure allows for simultaneous reconstruction of object and background, thus making it amenable to the new multi-thread / multi-processor architectures.

A computer system that segments a colon for a computed tomography colonography (CTC) is described. During operation, the computer system accesses imaging data having a spatial resolution. Then, the computer system identifies the colon lumen based on probabilities for different tissue classes in the imaging data. Moreover, the computer system segments the colon into subsegments based on an articulated object model that fits a tortuosity of the colon along a centerline of the colon, where the articulated object model includes values of an orthonormal basis set, curvature and torsion along the centerline, and where boundaries between subsegments are based on the curvature and die torsion. For example, a given boundary between a pair of subsegments may corresponds to or may be related to a minimum value of the curvature and a maximum value of the torsion over a length of the colon.

The invention relates to a random sampling analog circuit compressed sensing measurement and signal reconstruction method, which belongs to the field of electronic system test and fault diagnosis. Aiming at a fault signal having a sparsity distribution characteristic per se or in an orthogonal space in an output response of an analog circuit, a test node is selected according to a circuit topology structure, circuit output responses are randomly sampled under a distributed sensor test network, response signals are expressed in a sparse way on a transform domain by utilizing discrete orthonormal basis, compressed sensing measurement of the sparse signals is completed under observability matrix projection, and when the recovery rate of signal reconstruction by randomly compressed sampling points reaches more than 80 percent, the compressed measurement values of the circuit output responses are effective, can form a characteristic set and can be used for analog circuit fault diagnosis. The method solves the problems that the traditional analog signal sampling occupies a large number of hardware resources, large signal reconstruction calculated amount and the like; and the random sampling compressed sensing measurement method is utilized to improve the efficiency of electronic system testing.

A method for simplifying a passive electromagnetic transient-state simulation model of the large-scale power distribution network includes steps of dividing a large-scale intelligent power distribution system into an external system and a research system which are connected; respectively establishing an electromagnetic transient-state simulation model of the external system and the research system; setting a low-dimension system model order q; calculating A=-G<-1>C, R=G<-1>B according to a state-output equation model of the external system; selecting a corresponding basis calculating method according to the number of input quantity of the external system, solving a standard orthogonal basis V of Kq(A, R, q) of a krylov subspace of the q dimension; calculating Cq=V<T>CV, Gq=V<T>GV, Bq=V<T>B, Lq=V<T>L to obtain a low-dimension simplified system model shown in the description, utilizing a simplified reduced-order model y=Lq<T>xq of the external system to substitute for an original external system model, and simultaneously subjecting a detailed model of the research system to simulation calculation to obtain an internal detailed transient process of the research system. The method for simplifying the passive electromagnetic transient-state simulation model of the large-scale power distribution network has the advantages of high accuracy, good stability, simplicity in algorithm and proneness to realizing.

Systems and methods for robust recognition of machine-readable symbols from highly blurred or distorted images. An image signal representation of a machine-readable symbol element is transformed into a different space using one or more transform operations, which moves an n-dimensional vector of measured light intensities into another n-dimensional space. The types of transform operations may include blur robust orthonormal bases, such as the Discrete Sine Transform, the Discrete Cosine Transform, the Chebyshev Transform, and the Lagrange Transform. A trained classifier (e.g., an artificial intelligence machine learning algorithm) may be used to classify the transformed signal in the transformed space. The types of trainable classifiers that may be used include random forest classifiers, Mahalanobis classifiers, support vector machines, and classification or regression trees.

The present invention is directed to video coding / decoding and discloses a method for transforming to / from transform coefficients and residual pixel data in moving pictures by a set of semi-orthonormal basis vectors. The basis vectors are derived from conventional DCT or KTL matrixes, but relaxes to some extend the requirements for orthogonality, norm equality and element size limitation. In this way the present invention provides improved coding efficiency and lower complexity compared to previously used integer transforms.

Disclosed is a quantum image encryption method based on image association dissection. A quantum-state superposition and measurement principle is used to establish correlation between image pixels. An image is dissected into superposition of a series of characteristic sub-images. Conversion operations are performed on the characteristic sub-images stored in a complete binary-tree digit group through use of a random phase door and a quantum rotating door and then a plurality of superpositions are performed on all the images through quantum-state superposition so as to obtain a cipher image. Schmidt orthogonal dissection is performed on the quantum random phase door, the quantum rotating door, a coefficient matrix and the cipher image so as to obtain a standard orthogonal ground state which is then used as a key. The quantum image encryption method based on the image association dissection has a larger key space so that strong attacks can be resisted and combination of a quantum-mechanical theory and an image encryption technology is realized. The method has classical information theory security and quantum information theory security so that quantum image encryption surpasses restriction of classical image encryption and compared with classical images, quantum images are higher in security.

The invention relates to a compressive sensing imaging method for a synthetic aperture radar. The method includes: constructing a standard orthogonal basis for indicating radar echo signals by the aid of a Chirp Scaling operator, processing distance-vector and azimuth-vector two-dimensional data simultaneously by compressive sensing, and sparsely optimizing finally so as to obtain high-resolution radar images based on low data size. By the method, the data size is decreased in the distance vector and the azimuth vector firstly, then the standard orthogonal basis for indicating the radar echo signals is constructed by the aid of the Chirp Scaling operator, and two-dimensional sparse optimization is performed in the distance vector and the azimuth vector by means of compressive sensing so that the radar images are obtained. The data size can be decreased effectively under the circumstance of sparse observation scenes to lower requirements on data storage devices and transmission devices of the synthetic aperture radar, and the compressive sensing imaging method is of great practical significance to a practical system.

A method to perform secure boolean search over encrypted documents is disclosed. Each document is characterized by a set of keywords, all the keywords characterizing all the documents forming an index, the index being translated in an orthonormal basis in which each keyword of the index corresponds to one and only one vector of the orthonormal basis, each document being associated with a resultant vector in the span of the orthonormal basis, the resultant vectors corresponding to all the documents being stored in an encrypted search server. The method includes steps of receiving a search query from a querier; transforming the search query in one query matrix; and determining a general result based on the result of the multiplication between the query matrix and the resultant vectors.

The invention discloses a quick image super-resolution reconstruction method. The method comprises the steps of segmenting an image in a training image database into image blocks by utilizing a priori that natural image textures have similarity, and stretching the image blocks into vectors; then training a sparse code dictionary by utilizing the vectors; training the sparse code dictionary for corresponding high and low-resolution image data sets to obtain corresponding high and low-resolution dictionaries; and for a new input low-resolution image, calculating a sparse code by utilizing the low-resolution dictionary first and then multiplying the high-resolution dictionary by the code to obtain a high-resolution image. In a spares coefficient solving process, sparse coefficients of standard orthogonal basis are calculated for a fixed dictionary first, and for a new input sparse coefficient, the sparse coefficient is quickly solved in a manner of weighted summation of the sparse coefficients of the standard orthogonal basis. In an image block processing process, the image blocks are processed in parallel by adopting a CUDA technology, so that the processing time can be shortened to one ten-thousandth of conventional CPU calculation. The method is used for quickly obtaining a super-resolution image of a single image.

The invention provides a light-source optimization method adopting a compressed sensing technology. The light-source optimization method comprises the following steps: firstly selecting a group of standard orthogonal basis and enabling a light-source graphic to be sparse on the standard orthogonal basis; then expanding the light-source graphic on the standard orthogonal basis to obtain coefficient vector theta and constructing the SO optimization problem into an image restoration problem of solving a minimum L1 norm of theta under the linear-limitation condition, wherein the linear-limitation condition is as follows: imaging in photoresist on a plurality of observation points set at a wafer is consistent to a target graphic. Compared with the traditional SO algorithm adopting the conjugate gradient method, the SO method has the advantages that the operation efficiency can be more effectively improved, and the manufacturability of the optimized light source and a process window of a photoetching system can be improved; simultaneously, a vector imaging model is adopted for describing the imaging process of the photoetching system, the optimized light source not only is applicable to the condition of small NA, but also is applicable to the condition that NA is more than 0.6, and meets the requirement for the simulation precision of high-NA immersed type photoetching system.

A method for estimating the quantity of light received by a point M of a heterogeneous participating media, the light being emitted by a light environment. Thereafter, the method comprises estimating, for each point of a first set of points, first values representative of the light attenuation between the considered point and a first surface bounding the media along a plurality of particular directions of light emission, estimating first coefficients of projection by projection of the first values of reduction of light intensity in an orthonormal basis of spherical functions, estimating second values representative of the light attenuation between the point and a second surface along directions, the second surface comprising some of the points of the neighbourhood of the point, and estimating the quantity of light received by the point using first coefficients of projection and second values of reduction of light intensity.

The invention discloses a gas concentration detection method, which is characterized by adopting a KPCA (Kernel Principal Component Analysis) algorithm for identifying a large number, firstly building two mixed kernel functions, utilizing a vector method for building a kernel matrix, and utilizing kernel principle component analysis for calculating characteristic vectors of the kernel matrix, wherein the algorithm has higher recognition rate and higher arithmetic speed. The algorithm converts a KPCA process on a training set into a PCA process on a data set of coordinates of all kernel training samples under a group of basis through a group of standard orthogonal basis of a subspace expanded by the training samples in a characteristic space, meanwhile, carries out characteristic extraction on the training samples so as to effectively capture nonlinear characteristics of training data, and is widely interested and applied in mode recognition and regression analysis. During a solving process of KPCA, an M*M kernel matrix (M represents the number of the training samples) needs to be subjected to eigenvalue decomposition, so that when sample characteristics are extracted, only a sample and a kernel function for forming the group of Kidd samples need to be calculated, and an experimental result verifies that the algorithm is effective.

The invention discloses a safe and effective method for dynamically updating a wireless sensor network program, belonging to the field of wireless communication. The method comprises the following steps of: determining an euclidean space by a base station and dividing the euclidean space into a first subspace and a second subspace which are orthogonal to each other before deploying sensor nodes; randomly selecting a standard orthogonal basis of the first subspace, and randomly selecting a vector for each sensor node from the second subspace; determining a hash function, and selecting a same random number for all the sensor nodes as a cluster key; installing the vector, the cluster key and the hash function into each sensor node; broadcasting program updating message to a sensor network by the base station when the base station needs to update the programs on all the sensor nodes, and checking the validity of the message by each sensor node; if the program updating message is valid, receiving the program updating message by the sensor node and updating programs per se; otherwise, rejecting the program updating message.

The present invention discloses a non-divergence smoothing processing method for a velocity field, comprising: expanding three velocity components of a three-dimensional velocity field into a column vector velocity field according to a coordinate index; constructing a non-divergence smoothing standard orthogonal basis vector set; calculating a velocity field smoothing parameter; performing non-divergence smoothing on the column vector velocity field according to the non-divergence smoothing standard orthogonal basis vector set and the velocity field smoothing parameter obtained through calculation; and reducing the smoothed non-divergence column vector velocity field into a smooth three-dimensional velocity field with the same structure with an original velocity field. The present invention further discloses a non-divergence smoothing processing apparatus for the velocity field. According to the present invention, not only can a high-frequency noise error be eliminated, but also the smoothed velocity field can meet a continuous equation of incompressible fluid.

A video coding / decoding system, method and computer program product employ an integer transform matrix for transforming to / from transform coefficients and residual pixel data in moving pictures by a set of semi-orthonormal basis vectors. The basis vectors are derived from conventional DCT or KTL matrixes, but relaxes to some extent the requirements for orthogonality, norm equality and element size limitation. In this way improved coding efficiency and lower complexity compared to previously used integer transforms are possible.

The invention discloses a coal mine gas prediction method. The KPCA algorithm is used for identifying a large number. Firstly, two hybrid kernel functions are constructed, a kernel matrix is constructed through a vector method, and a feature vector of the kernel matrix is calculated through kernel principal component analysis, wherein the algorithm is high in recognition rate and calculation speed; according to the algorithm, by means of a group of standard orthonormal bases of a subspace spanned by training samples in a feature space, the KPCA process on a training set is converted into the PCA process in which coordinates of all the kernel training samples under the group of bases are datasets, meanwhile, feature extraction is conducted on the training samples so that nonlinear features of training data can be effectively captured and widely paid attention to and applied in mode recognition and regression analysis. In the KPCA solving process, feature values need to decompose the M*M kernel matrix (M refers to the number of the training samples), when feature extraction is conducted on one sample, people only need to calculate the sample and the kernel functions between the samples forming the group of bases, and experiment results verify that the algorithm is effective.

The invention discloses a gas concentration detection method. According to the method, a KPCA algorithm is used to identify the "large number", wherein two mixed kernel functions are firstly constructed, a kernel matrix is constructed by using a vector method, and the characteristic vector of the kernel matrix is analyzed and calculated by using the kernel main component; the algorithm has characteristics of high recognition rate and high operation speed; according to the algorithm, the KPCA process on a training set is converted into the PCA process of all kernel training samples having the coordinate of the data set at the group basis through a group of training sample standard orthogonal basis of the sub-space formed in the character space, while the training sample is subjected to character extraction, such that the non-linear characteristics of the training data can be effectively captured, and the wide attention and application is provided in pattern recognition and regression analysis. According to the present invention, during the calculation process of the KPCA, the character value is required to decompose the M*M kernel matrix (M represents the number of the training samples); and when the sample character is extracted, only the kernel function between the sample and the samples for forming the group basis is required to be calculated, and the experiment results verify that the algorithm is effective.

The invention discloses a method for recovering phase distribution of phase shift interference figures and a method for obtaining phase shift between the two figures. The method for obtaining the phase shift between the two figures comprises writing each interference figure in the form of a matrix and serving each interference figure as a column vector matrix; obtaining a background component through a low-pass filter; subtracting each column vector formed by the corresponding phase shift interference figure from the background component to obtain a pair of vectors and forming a subspace by the two vectors to achieve the vectorization of the interference figures, wherein an included angle is formed between the vectors; obtaining an orthogonal basis of the subspace through a basis in the subspace by a Schmidt orthogonalization method, further obtaining a corresponding standard orthogonal basis and obtaining an included angle formed between two original signals through the obtained standard orthogonal basis. According to the method for recovering the phase distribution of the phase shift interference figures and obtaining the phase shift between every two figures, the phase distribution between the two phase shift interference figures can be obtained and the changes of the phase shift can be further obtained through the series of phase shift interference figures.

A device to apply a digital fingerprint to a digital signal comprises a means of intercepting or acquiring a signal, a storage element and a processor for executing computer implemented programme code components in the storage element to effect the methods. The methods include transforming a plurality of signal samples onto a discrete orthonormal basis and ranking the transformed samples according to their magnitude. The n largest principal components of the ranked transformed samples are optionally permuted to generate a re-ordered set of principal components, which are then altered by a marking angle. The marked principal components and unmarked non-principal components are converted and combined and applying an inverse of the transform function to the combined principal and non-principal components to generate a fingerprinted digital signal. Methods to prepare the signal for marking, recover the digital fingerprint and verify the distributor and / or recipients of the signal are also disclosed.

A video compression framework based on parametric object and background compression is proposed. At the encoder, an embodiment detects objects and segments frames into regions corresponding to the foreground object and the background. The object and the background are individually encoded using separate parametric coding techniques. While the object is encoded using the projection coefficients to the orthonormal basis of the learnt subspace (used for appearance based object tracking), the background is characterized using an auto-regressive (AR) process model. An advantage of the proposed schemes is that the decoder structure allows for simultaneous reconstruction of object and background, thus making it amenable to the new multi-thread / multi-processor architectures.

The invention provides a signal subspace technology based single-channel spaceborne SAR moving target detection method, which can effectively suppress residual clutters and reduce the false alarm probability. The method comprises the steps of: first, carrying out singular value decomposition on a reference image matrix; second, solving a standard orthogonal basis of a reference image clutter subspace; third, solving the projection of an image to be detected in the reference image clutter subspace, and solving the projection of an f2 (x, y) in an f1 (x, y) subspace according to the standard orthogonal basis of the reference image clutter subspace; fourth, solving an image of the image to be detected after clutter suppression; and fifth, normalizing the image after clutter suppression by using an average value of a reference image block.

The invention relates to a method for estimating the quantity of light scattered by a heterogeneous participating media, the light being transmitted by a light environment comprising a plurality of light sources. In order to improve the display while minimising the required calculation time, the method comprises steps for:estimation of first projection coefficients in an orthonormal basis of spherical functions, said first projection coefficients being representative of the reduction of light intensity in a point of said media, said estimation of first projection coefficients being carried out for each point of a first set of points of said media, andestimation of the quantity of light scattered by said media, according to at least one scattering direction of the light, using said first estimated projection coefficients.

Method for determining best and worst cases for values of model parameters such as porosity and shale volume fraction generated by non-unique matrix inversion of physical data such as seismic reflection amplitudes. The matrix is diagonalized, and then orthonormal basis vectors associated with insignificant diagonal elements are used to generate upper and lower bounds on the solution. Best and worst case solutions are determined as linear combinations of the null basis vectors, where the expansion coefficients are determined by making a best fit to the upper and lower bounds.

The invention discloses an interference elimination method based on Givens conversion block diagonilization precoding. The method specifically comprises the first step of estimating a channel matrix which participates the precoding; the second step of conducting QR resolving on the channel matrix of a united user after conjugating transposition; the third step of resolving pseudo-inverse of the channel matrix of the united user; the fourth step of constructing the first half of precoding matrix; the fifth step of constructing the second half of the precoding matrix; the sixth step of obtaining the precoding matrix of final union. According to the interference elimination method based on the Givens conversion block diagonilization precoding, orthogonal matrix with better orthogonality is obtained, thus more accurate standard orthogonal basis of user channel matrix null space is obtained, and finally the needed precoding matrix of the final union is obtained. The interference elimination method based on the Givens conversion block diagonalization precoding can be used in a multi-user communication system to eliminate interference among users and improve the quality of signals received by the users and the stability of the system.

The invention discloses a processing method for converting triaxial acceleration data collected by an acceleration sensor in a mobile phone to a geodetic reference coordinate system from a mobile phone coordinate system. The reference coordinate system is based on a standard orthogonal basis and completely independent of the mobile phone coordinate system; the converted triaxial acceleration datahas a clear physical meaning and can reflect a motion trend of a mobile phone user at each axis more accurately; the X and Y axis data records a motion of the user in a horizontal direction (front / back, left / right); and the Z axis data records a motion of the user in a vertical direction (up / down). Therefore, the influence of a mobile phone orientation on the data collected by the internal acceleration sensor is effectively reduced, so that the correct rate of identifying behaviors of the mobile phone user is remarkably increased; and the method has wide application values in mobile terminal-based context awareness.

A method includes capturing multivariate time-series data comprising two or more data sets from a system captured over a past time period and a present time period, applying at least two sliding time windows to the multivariate time-series data in determining respective data matrices, computing an orthonormal matrix for each of the data matrices, wherein the orthonormal matrix is a signature of fluctuation patterns of a respective data matrix, computing a difference between at least two of the data sets in the past and the present time periods through the orthonormal matrices, and detecting a fault in at least one of the systems by comparing the difference to a threshold.