114 results about "Independent component analysis algorithm" patented technology

The invention discloses a face recognition method by an independent component analysis based on a multi-scale total variational derivative image, which belongs to the face recognition technical field; the method is as follows: a contrast gradient is strengthened; TV-L<1> is used to carry out scale decomposition to a face image to obtain a large-scale image comprising a skeleton contour and muscle information and a small-scale image comprising the details of mouth, eyes and nose; quotient balance is carried out to the small-scale image to obtain the feature of unchanged illumination; feature fusion technology is selected to fuse the features of large scale and unchanged illumination into a new face image; Gabor is used to analyze and extract the features of the new face image in a specific scale and direction to generate a multi-scale Gabor face; the eigenvectors of all the samples are extracted by an information maximization independent component analysis algorithm; the similarity of the eigenvectors of the face which is to be treated with recognizing is calculated by the eigenvectors of the known face; according to the similarity, the eigenvectors are sorted to acquire a final recognition result. The face recognition method achieves high recognition rate and strong robustness to illumination, expression, make-up and other external interference.

Methods and apparatus for signal processing are disclosed. Source separation can be performed to extract source signals from mixtures of source signals by way of independent component analysis. Source separation described herein involves mixed multivariate probability density functions that are mixtures of component density functions having different parameters corresponding to frequency components of different sources, different time segments, or some combination thereof.

The invention relates to a wind-speed time series forecasting method for a wind power station. The wind-speed time series forecasting method is characterized by comprising the following steps that: a wind-speed collecting instrument is used for data of wind speed of same area once per hour, and the collected original wind speed data is organized to form a wind-speed time series for analyzing and forecasting; a rapid independent component analysis algorithm is utilized for carrying out multiple-scale decomposition on the wind-speed time series, so as to obtain a plurality of independent components; the delay time and the embedding dimension of the independent components are calculated, and a phase-space reconfiguration theory is adopted for carrying out phase space reconfiguration on the independent components; a least squares support vector machine regression model is utilized for carrying out modeling forecasting on the independent components after phase space reconfiguration; and the forecasting results are superposed to obtain the final forecasting result of the wind-speed time series. The method is scientific, reasonable, accurate and reliable in wind-speed time series forecasting, and has strong adaptivity.

The invention relates to a harmonic current estimation method under the condition of unknown harmonic impedance, and the method comprises the steps: taking a harmonic voltage as a measurement quantity and taking a harmonic current as a state quantity according to the statistic independence and super-Gaussian distribution characteristics of a harmonic current quick fluctuation quantity, and building a harmonic state estimation model comprising measurement noise; enabling the noise and impedance parameters in an ICA model to serve as unknown variables, enabling the harmonic current to serve as a hidden variable, and obtaining a harmonic current optimal solution which cannot be affected by noise through employing the variational Bayesian learning capability for unknown variables; neglecting linear load admittance according to the characteristic that the linear load admittance is far less than system side admittance, eliciting a harmonic current amplitude proportion coefficient, and determining the current amplitude. Compared with the prior art, the method can effectively eliminate the uncertainty of an independent component analysis algorithm and restores the amplitude of the harmonic current with no need of harmonic impedance parameters, obtains an estimated harmonic current, can effectively reduce the impact on the estimation result from measurement noise, and is stronger in robustness.

The invention provides a front car identification method based on monocular vision. The method includes the steps that (1), an original image is collected from a vehicle-mounted camera, the edge of the image is extracted according to a Canny edge extraction method, influence of noise points is eliminated through morphological filter, projection is carried out in the horizontal direction, and an area of interest of a front car is obtained according to projection characteristics; (2), a shadow area at the car bottom is extracted and judged according to the geometrical shape of the shadow at the car bottom, edge characteristics are overlaid, and a car area is judged; (3), graying, normalization and binary tree complex wavelet transformation are carried out on small color images of candidate car areas of different shapes, and characteristic vectors are obtained; (4), the number of dimensions of the characteristic vectors is decreased through a two-dimension independent component analysis algorithm, the characteristic vectors are fed into a support vector machine based on a radial basis function kernel to be classified, and it is judged that whether the candidate car areas are the car area. Cars on the road ahead are detected accurately, and real-time and reliable road condition information can be supplied for unmanned cars.

A portable stimulating, awaking and evaluating system for disturbance of consciousness comprises an EEG acquisition device, a signal preprocessing device which is connected with the EEG acquisition device and used for extracting an independent source component from a first EEG using an independent component analysis algorithm, a digital filter device which is connected with a signal preprocessing device and used for carrying out digital filter process on a second EEG using an empirical mode decomposition algorithm so as to extract an effective third EEG, a feature extracting device which is connected with the digital filter device and used for extracting frequency spectrum features, the sample entropy and the approximation entropy, a classification and a recognition device which is connected with the feature extracting device and used for classifying and recognizing disturbance of consciousness using a support vector machine according to the spectrum features, the sample entropy and the approximation entropy, and an output device which is used for outputting a stimulating signal through a stimulator. The portable stimulating, awaking and evaluating system for disturbance of consciousness is simple in structure and good in portability, and the portable stimulating, awaking and evaluating system for disturbance of consciousness can be used for families and communities using stimulation therapy on disturbance of consciousness.

The invention relates to the technical field of independent component analysis (ICA) and the field of fusion of multiple seismic attributes, and provides a method for fusing multiple seismic attributes on the basis of fast independent component analysis (FICA). The scheme includes that each attribute participating in fusion is divided into attribute blocks with identical sizes, the quantities of the attribute blocks of the attributes are identical, a certain quantity of blocks are selected from the attribute blocks and are processed according to an FICA principle to obtain a separation matrix and a reciprocal hybrid matrix of the separation matrix, and all the blocks are mapped to an ICA domain by the separation matrix; the corresponding blocks of the attributes are fused in the ICA domain according to fusion rules, and finally a fusion result of the ICA domain is mapped to a spatial domain to obtain a fusion result; and the fusion result is beneficial to analyzing complicated stratum information and improving reservoir prediction precision. The method can be widely applied to seismic attribute analysis, comprehensive interpretation, seismic reservoir prediction and lithological character and fluid identification.

The invention discloses a supervised multimodal brain image fusion method comprising the following steps: S1, calculating the characteristic of each modal; S2, matrixing and normalizing the characteristics of the modals; S3, reducing the dimension of the modal characteristics using a singular value decomposition algorithm; S4, based on the dimension-reduced modal characteristics obtained in S3, maximizing the sum of squares of correlation between the typical variables of the modals and between the typical variables and prior information, and performing an iterative cycle until convergence; and S5, connecting the modal components obtained in S4 in series, and calculating out an independent component and a mixed matrix of each modal significantly associated with the prior information using a joint independent component analysis algorithm, thus realizing supervised multimodal brain image fusion. According to the embodiment of the invention, the method has good robustness and can reveal the physiological and pathological mechanism of complex brain disease cognitive impairment.

The invention discloses a lead selection method for emotional electroencephalogram based on independent component analysis. A multi-lead emotional electroencephalogram is utilized for conducting filtering processing, ICA analysis is conducted on data after filtering, spatial filter sets corresponding to different emotional task backgrounds are constructed, and linear projection is conducted; spatial feature parameters of a full-lead emotional signal are acquired, and then the lead selection method is utilized for selecting the optimum lead set of a subject. According to the method, higher recognition accuracy rate is acquired, it is achieved that different subjects can automatically select emotion-related independent components, and compared with a mode of extracting the independent component of a full-passage, the independent component of the optimum lead position not only can reduce the time complexity of an algorithm but also can accurately describe the real conditions of an emotion-related independent source and effectively inhibit the components which are not related with emotional signals and the interference caused by external noises.

The invention discloses a self-adaptation independent component analysis method for harmonic wave impedance estimation, relates to the field of power systems and aims to solve the problems that in theprior art a conventional power system harmonic wave impedance estimation method is low in precision, a novel rapid independent component analysis algorithm is low in convergence performance and source signals cannot be all recovered with an optimal effect. The method comprises the following steps: firstly, sampling a harmonic wave voltage and harmonic wave current at a public coupling point; secondly, carrying out centralization and whitening processing on a harmonic wave voltage sampling value and a harmonic wave current sampling value as known mixed signals; thirdly, with the criterion of negentropy maximization, carrying out iteration so as to obtain a primary separation matrix; finally carrying out self-adaptation selection on nonlinear functions, and carry outing progressivity analysis till an optimal nonlinear function, and calculating a harmonic wave impedance matrix, thereby finally obtaining harmonic wave impedance. By adopting the method, limits of conventional methods are overcome, meanwhile defects of rapid independent component analysis are made up, and the harmonic wave impedance estimation precision can be improved.

The invention discloses a non-Gaussian process monitoring method based on multi-variable block interleaving correlation elimination, and aims to take interleaving correlation into a distributed process modeling and process monitoring, thereby applying more reliable and effective distributed non-Gaussian process monitoring. The method comprises the steps of dividing all measuring variables into a plurality of variable sub-blocks according to ownership of the variable which is measured by each production unit; secondly, eliminating interleaving correlation information between each variable sub-block and other variable sub-block by means of a regression model; and finally, performing modeling based on an independent component analysis algorithm and non-Gaussian process monitoring by means ofthe error after interleaving correlation elimination. Compared with a traditional method, the method of the invention is mainly advantageous in that the regression model is utilized for taking the interleaving correlation between different multi-variable sub-blocks is considered, and the error after interleaving correlation elimination is used as a new monitoring object.

Provided is an apparatus and method for separating music and voice using an independent component analysis method for a two-dimensional forward network. The apparatus of separating music and voice can separate voice signal and a music signal, each of which are independently recorded, from a mixed signal, in a short convergence time by using the independent component analysis method, which estimates a signal mixing process according to a difference in record positions of sensors. Thus, users can easily select accompaniment from their own compact discs (CDs), digital video discs (DVDs), or audio cassette tapes, or FM radio, and listen to music of improved quality in real time. Accordingly, the users can just enjoy the music or sing along. Furthermore, since the independent component analysis method in the apparatus of separating music and voice is simple and time taken to perform the method is not long, the method can be easily used in a digital signal processor (DSP) chip, a microprocessor, or the like.

The invention relates to a periodic long-short code direct sequence spread spectrum code division multiple access signal multi-pseudo-code estimation method. The existing direct sequence spread spectrum code division multiple access signal pseudo code blind estimation technology cannot be applied to periodic long-short code direct sequence spread spectrum code division multiple access signals adopting short code spread spectrum and long code scrambling. The method comprises the steps of: constructing periodic long-short code direct sequence spread spectrum code division multiple access signals with complicated structures into a deficit matrix model of short code direct sequence spread spectrum code division multiple access signals, and modeling compound code matrix estimation as a blind source signal separation problem; applying a matrix filling theory to the compound code matrix estimation, and estimating user compound code sequences based on a singular value threshold algorithm and a fast independent component analysis algorithm; and finally proposing a delay triple correlation algorithm by utilizing a displacement stacking feature of an m sequence, and estimating long-short pseudo code sequences contained in the user compound code sequences. The periodic long-short code direct sequence spread spectrum code division multiple access signal multi-pseudo-code estimation method fully utilizes the matrix filling mathematic model and the triple correlation peak feature of the m sequence, and achieves blind estimation of the user compound code sequences, a long scrambling code sequence and a short spread spectrum code sequence of the signals.

The invention discloses a signal separation method based on particle swarm optimization. The method includes the steps: (1) inputting observation signals; (2) building an independent component analysis algorithm optimization model by taking a difference value between simple products of a minimum separated signal joint probability and a marginal probability as an optimization target; (3) estimating the number of source signals according to a singular value decomposition method and determining the number of optimization variables according to the number of the source signals; (4) calculating correlation coefficients of the observation signals, and determining the value range of the optimization variables; (5) optimizing separation matrixes by the aid of a particle swarm optimization algorithm; (6) taking particles with optimal fitness in the last generation population after optimization as an optimal separation matrix and multiplying the optimal separation matrix with the mixed signals to obtain the optimal separation signals. The method based on independent component analysis of the particle swarm optimization has universal applicability and effectively solves various blind source separation problems.

The invention belongs to the field of image processing technology and discloses a multi-shape-prior level set independent component analysis method and an image partitioning system. The method comprises the steps that a to-be-partitioned image and shape priors are input; curve initialization is performed; shape prior alignment is performed; the aligned shape priors are coded through a level set function; a shape prior matrix is formed; independent component analysis is used to perform dimension reduction; the current level set function is projected to a low-dimension space; probability distribution of the shape priors is estimated, shape driving energy items are constructed and combined with data driving energy items, and an energy function is formed; and the energy function is minimized,curve evolution is driven, and a partitioning result is obtained. Through the method and the system, high-dimension redundant features of the shape priors can be eliminated, therefore, distribution ofthe shape priors can be subjected to more accurate statistical analysis, more effective shape constraint can be formed, and finally an accurate partitioning result can be obtained.

The invention discloses a satellite-borne AIS conflicting signal separation method based on improved independent component analysis. The method comprises the following steps: sampling by an analog-to-digital converter to form N channel receiving data; performing centralization processing and whitening processing on N paths of received observation signals; obtaining N paths of separation signals from the matrix preprocessed by the observation signals by using an improved ICA method based on the cuckoo algorithm and the Newton iterative algorithm; performing digital down-conversion, matched filtering and whitening filtering on the obtained N paths of separation signals; and decoding the N paths of separation signals after the whitening filtering by using the viterbi algorithm to obtain N AIS data frames. According to the satellite-borne AIS conflicting signal separation method disclosed by the invention, no frequency deviation, phase shift or amplitude estimation is required, a satellite-borne AIS receiver structure is simplified, and the satellite-borne load is relieved.

The invention discloses a method and device for diagnosing an on-load tap switch. The method for diagnosing an on-load tap switch comprises the steps of acquiring a vibration signal of the switch; extracting an independent component vector from the vibration signal; calculating the statistic and a squared prediction error of the independent component vector; judging whether the statistic and the squared prediction error are greater than corresponding preset confidence limits or not; when the statistic and the squared prediction error are greater than the corresponding preset confidence limits, projecting the independent component vector on multiple sections of a high-dimensional space, and obtaining independent component vector parameters of the vibration signal are acquired on the projection plane; and determining whether the switch has a fault or not according to the independent component vector parameters. According to the invention, the independent component vector is extracted from the vibration signal, and comparison with data parameters of a normal operating state is performed by combining an independent component analysis algorithm and a support vector machine classification algorithm, so that the fault is recognized quickly, the time for diagnosing the fault online is shortened, and the efficiency of fault diagnosis is improved.

The invention discloses a nonparametric estimation independent component analysis (ICA)-based multiple input multiple output-orthogonal frequency division multiplexing (MIMO-OFDM) system blind deconvolution method. The method comprises the following steps of: combining nonparametric estimation ICA and MIMO-OFDM; arranging a signal preprocessing module at a transmitting end of the MIMO-OFDM system to perform non-redundant linear precoding on a frequency domain signal mapped to a subcarrier so as to eliminate fuzziness of an independent component analysis algorithm in reconstructed signal sorting and energy; and arranging a nonparametric ICA module at a receiving end of the system to perform channel estimation so as to restore a source signal. By the method, frequency spectrum resource is saved, the amount of calculation is reduced, the operational speed is increased and a time-varying channel can be estimated well.

The invention relates to a microphone array-orientated acoustic echo cancellation method. According to the method, a microphone array acquires the data of near-end signals and echo signals in real time; the channel signals of one microphone are selected to construct an initialization model according to an adaptive filtering idea and an independent component analysis algorithm; an echo cancellationmodel is established according to the initialization model to perform echo cancellation on the single-channel signals of the microphone; and the echo cancellation models of the other microphones in the microphone array are constructed according to the initial echo cancellation model and a single-channel filter coefficient, so that the echo cancellation models can be adopted to perform echo cancellation. With the microphone array-orientated acoustic echo cancellation method of the invention, acoustic echo cancellation can be performed on the multi-channel microphone array, so that effective data processing can be provided, and a speech identification rate can be improved. According to a device provided by the invention, acoustic sensors are distributed at key positions on a ring-shaped sensor array so as to measure the multi-channel sound record signals and echo signals of a system and perform echo cancellation on the multi-channel sound record signals and the echo signals, so that theecho cancelled signals of the microphone array can be obtained, and therefore, a speech identification rate can be improved.

The invention relates to a multi-harmonic source identification method for a power distribution network. The method comprises the following steps of (1) placing a phasor measurement unit (PMU) in thepower distribution network; (2) performing optimization configuration of the PMU on the power distribution network: determining a mathematic model of PMU optimization configuration; solving the mathematic model of the PMU optimization configuration by utilizing a binary imperialist competitive algorithm to obtain an optimization configuration scheme of the PMU; (3) obtaining harmonic voltage Vh data in a period of time at a node for configuring the PMU in the power distribution network; (4) obtaining a fast change component Vh.fast in a harmonic voltage Vh by using a mobile filter; (5) preprocessing Vh, namely, performing centralization processing and whitening processing; and (6) applying the fast change component Vh.fast to an improved independent component analysis algorithm, determining a separation matrix, and enabling the separation matrix to act on the harmonic voltage Vh for separation to obtain harmonic currents Ih injected into the power distribution network. The method has the characteristics of systematization and accurate estimation result, and can be used for the multi-harmonic source identification process under unknown harmonic impedance in the actual power distribution network.

The invention puts forward a simple target identification method based on channel state information and a support vector machine, does not need to build special hardware facilities, fully utilizes an existing wireless network and can realize a simple target identification function by a common commercial router. After CSI (Channel State Information) original data is obtained, firstly, DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is adopted to carry out clustering on subcarrier data in a channel to carry out denoising, and then, a weight-based moving average algorithm is adopted to smoothen data subjected to the denoising. After the data is preprocessed, a principal component analysis algorithm is adopted to carry out feature value extraction on the data. The data subjected to preprocessing and feature extraction can more accurately reflect the main change of a signal, and in addition, a dimension is greatly lowered so as to be favorable for improving target identification accuracy and lowering calculation complexity. By use of the method, be means of a SVM (Support Vector Machine) multi-classification algorithm based on a one-against-one strategy, a statistic model under a nonlinear dependence relationship between a target object and a signal fingerprint is obtained so as to achieve a purpose of simple target identification.

The invention discloses a motor imagery electrocorticogram (EEG) signal classification method based on independent component analysis, which comprises the following steps: S1, collecting the EEG signals and preprocessing the EEG signals, and randomly dividing the preprocessed EEG signals into a training set and a test set; S2, sequentially selecting the training set data to carry out independent component analysis and calculation on the single test sample data, and automatically identifying and acquiring the motion related component based on the spatial distribution pattern of the source; S3:classification and recognition of motor imagery based on zero training classifier; S4: optimizing the selection of leads by using the training set data, substituting the optimized leads into the testset, and looping the steps S2 and S3 to obtain the final classification recognition rate. The invention can reduce the spatial model matching problem caused by the difference between the collected EEGdata, and has high recognition accuracy to the motor imagery EEG signals.

The invention relates to a face identification method based on an independent component analysis network. The method comprises the following steps: 1) carrying out cutting, aligning and normalizationpretreatment on input face images; 2) carrying out filtering through a group of trained ICA filters to obtain a group of mapping images; 3) carrying out non-linear and pooling processing on each mapping image to obtain a more efficient feature mapping image; 4) carrying out block LBP coding on each mapping image, and then, stringing the obtained local features to obtain feature expression; and 5)carrying out WPCA dimension reduction on the feature expression, and finally, carrying out identification verification on the two pieces face images through a cosine similarity measurement method. Thetrained ICA filters are applied to a CNN to form a single network; multi-scale information can be obtained based on different sensing areas of the ICA filters; and the method can ensure higher recognition rate in face recognition, and meanwhile, reduces calculation amount effectively and is convenient for popularization and application.

The invention discloses an independent components analysis (ICA) blind signal separation method and an ICA blind signal separation system based on a smoothing function and Parzen window estimation. A new ICA blind signal separation method is provided on the basis of Parzen window estimation technology assisted by a newly constructed smoothing function, so that a probability density function and a hybrid matrix of a source signal are estimated, and an unknown blind source signal is effectively separated out. A corresponding separation system comprises a signal receiving module, a single preprocessing module, a NewICA reconstruction source signal module and a subsequent processing module which are connected in turn. An effective ICA blind signal separation method and an effective ICA blind signal separation system with small error and high signal to interference ratio are provided.

The invention discloses an automobile sensor fault detection method based on independent component analysis and a sparse denoising auto-encoder. The method comprises the following steps: firstly, obtaining non-Gaussian information in process data by using independent component analysis to obtain independent component components, and extracting main independent components by using a sparse denoising auto-encoder to calculate an I2 index; and obtaining the Gaussian information of the operation data in the residual error space by using the sparse denoising auto-encoder to calculate the H2 index;finally, analyzing a fault detection effect by using fault false alarm rate (FAR) and a false detection rate (MDR) index. Compared with other methods, independent component analysis and the sparse noise reduction auto-encoder are combined, the sparse noise reduction auto-encoder is used in the non-Gaussian part to extract a principal element, and unnecessary signal interference is removed; gauss information in the data is extracted in a residual space by using a sparse denoising auto-encoder, so that the robustness of a process monitoring system is improved, the processing capability of nonlinear data is enhanced, and the accuracy of fault diagnosis is improved.

The invention discloses a method for improving the detection accuracy of blood oxygen saturation. The method comprises the following steps of: first, acquiring an original blood oxygen signal and performing band-pass filter on the original signal; then, calculating the blood oxygen signal subjected to the band-pass filter by an information maximum independence-based independent component analysis algorithm to obtain a blood oxygen independent component and an interference independent component and estimating a separated matrix W and a mixed matrix A; identifying and removing the interference independent component and acquiring two paths of clean blood oxygen signals; and finally calculating the blood oxygen saturation by using the clean blood oxygen signals. By the method for improving the detection accuracy of the blood oxygen saturation in the technical scheme of the invention, the interference caused by noise during calculation of a rate value can be effectively reduced, the aim of improving the detection accuracy of the blood oxygen saturation can be fulfilled by setting and judging an accuracy value; and the method has the advantages of extracting a double-wavelength blood oxygen signal acquired under different environments on line and ensuring more stable calculation accuracy, along with low operation complexity.

The invention discloses an X-ray medical image objective reconstruction method based on independent component analysis. In order to solve the problems that a traditional medical X-ray image is high in noise and poor in layering, and organs and tissue overlap, image denoising and objective extraction are carried out by combining a multi-energy-spectrum X-ray imaging technology and an independent component analysis algorithm. The method comprises the steps that first, denoising preprocessing is carried out on a medical image so that the preconditions for objective separation through independent component analysis can be met; second, the thickness value of an aliasing organ in each pixel is obtained according to an X-ray attenuation energy matrix of organ and tissue; third, the independent component analysis algorithm is used for adjusting the frequency of convergence and the size of signal scales according to the thickness value of the aliasing organ, a convergence matrix is obtained, the image of each organ is separated, contrast ratio correction is carried out according to an subjective vision standard, an area of interest and marginal information stand out, and the visual and clear image applicable to medical analysis are obtained.