60 results about "Kernel regression" patented technology

A method of image processing using kernel regression is provided. An image gradient is estimated from original data that is analyzed for local structures by computing a scaling parameter, a rotation parameter and an elongation parameter using singular value decomposition on local gradients of the estimated gradients locally to provide steering matrices. A steering kernel regression having steering matrices is applied to the original data to provide a reconstructed image and new image gradients. The new gradients are analyzed using singular value decomposition to provide new steering matrices. The steering kernel regression with the new steering matrices is applied to the noisy data to provide a new reconstructed image and further new gradients. The last two steps are repeated up to ten iterations to denoise the original noisy data and improve the local image structure.

A method for determining the future operational condition of an object includes obtaining reference data that indicates the normal operational state of the object, and obtaining input pattern arrays. Each input pattern array has a plurality of input vectors, while each input vector represents a time point and has input values representing a plurality of parameters indicating the current condition of the object. At least one processor generates estimate values based on a calculation that uses an input pattern array and the reference data to determine a similarity measure between the input values and reference data. The estimate values, in the form of an estimate matrix, include at least one estimate vector of inferred estimate values, and represents at least one time point that is not represented by the input vectors. The inferred estimate values are used to determine a future condition of the object.

A system for monitoring in real-time the health of equipment or the performance of a process utilizing a universal modeling technique that generates estimates of parameters for gauging early indications of anomalies. A kernel regression model such as the Nadaraya-Watson may be used, and may be in autoassociative form. Kernel optimization is automatically provided. A support vector regression can be substituted for the kernel regression.

A monitoring system for determining the condition of an object includes an empirical model with reference data that indicates the normal operational state of the object, and input pattern arrays. Each input pattern array has a plurality of time-ordered input vectors, while each input vector has input values representing a plurality of parameters indicating the current condition of the object. The empirical model generates estimate values based on a calculation that uses an input pattern array and the reference data to determine a similarity measure between the input values and reference data. A differencing module compares the estimates values to the corresponding input values so that resulting values from the comparison can be used to determine the condition of the object.

A method for monitoring the condition of an object includes obtaining reference data that indicates the normal operational state of the object, and obtaining input pattern arrays. Each input pattern array has a plurality of time-ordered input vectors, while each input vector has input values representing a plurality of parameters indicating the current condition of the object. Then at least one processor generates estimate values based on a calculation that uses an input pattern array and the reference data to determine a similarity measure between the input values and reference data. The estimate values are compared to the corresponding input values so that resulting values from the comparison can be used to determine the condition of the object.

Disclosed is a prediction method for monitoring performance of power plant instruments. The prediction method extracts a principal component of an instrument signal, obtains an optimized constant of a SVR model through a response surface methodology using data for optimization, and trains a model using training data. Therefore, compared to an existing Kernel regression method, accuracy for calculating a prediction value can be improved.

A method and apparatus for reconstructing a high-resolution image based on multiple low-resolution images are disclosed. The method and apparatus incorporating an embodiment according to the present invention reconstructs the high-resolution image based on a kernel regression method using a modified kernel function. The kernel function takes into consideration of registration reliability of regression residue and rotational motion within the multiple low-resolution images. The registration reliability adjusts weighting on the regression residues according to local gradient estimated between neighboring values. Furthermore, multi-scale regression residue is used to alleviate impact of noise.

A fast super-resolution reconstruction method is based on kernel regression for single frame images, and the invention relates to an image super-resolution reconstruction method, which overcomes the shortages that the existing super-resolution reconstruction method of the single frame images of kernel regression is large in calculated amount and long in consuming time. The invention comprises steps as follows: mapping the pixels on the low-resolution image to high-resolution grids, confirming the pixels which are needed to be estimated and classifying the pixels into two types, confirming quadrate neighborhood pixel aggregates of each pixel which is needed to be estimated in the first type and introducing the aggregates to a kernel regression equation to calculate the pixel value, conforming rhombic neighborhood pixel aggregates of the pixels needed to be estimated in the second type and introducing the aggregates to the kernel regression equation to calculate the pixel value and outputting images when all the pixels which are needed to be estimated are value-assigned. The invention introduces two-dimension nonlinearity kernel regression to estimate interpolation points, employs local neighborhood operation to replace whole image processing, and employs immediate updating strategy, thereby realizing the super-resolution reconstruction of the single frame images.

The invention discloses a kernel regression-based image compression sensing reconstruction method, which mainly solves the problem of reduced quality of a reconstructed image caused by mutually independent reconstruction of each image block and lack of considering linkage between the image blocks existing in the conventional method. The method comprises the following steps of: partitioning an input scene image; performing preliminary reconstruction on the image blocks by using an orthogonal matching pursuit (OMP) algorithm; then performing a kernel regression method on the image to obtain a local gray matrix of the image small blocks; weighing by using neighborhood image blocks to obtain a non-local gray matrix of the image small blocks; and finally, solving the final reconstruction imagesmall blocks through least square by using the local gray matrix and the non-local gray matrix of the image small blocks, and repeating the operation on all the image small blocks to obtain the finalreconstructed image. In the invention, both the reconstruction effects of various natural images and cartoon images can be improved under different sampling rates; and the method can be used for compressing high-resolution recovery or reconstruction of various low-resolution images under observation.

The invention discloses a method for removing image noise based on kernel regression total variation, and belongs to the field of computer vision. The method comprises the steps: (1), enabling a kernel regression method to be extended in a bounded total variation space, and constructing a regularization item comprising noise image local priori knowledge; (2), constructing a regularization item of non-local priori knowledge through employing non-local similar information in an image, and building an image noise reduction model; (3), carrying out the quick solving of the constructed model through employing a split Bregman method, and obtaining a preliminary estimated image; (4), carrying out the residual error iteration optimization of the preliminary estimated image, and achieving the optimal image noise removing effect. The method keeps the detail and texture information of the image well while maintaining an image structure.

Embodiments are directed to mobile localization, and more specifically, but not exclusively, to tracking mobile devices. Embodiments include methods that consider probability kernels with distance-like metrics between distributions. Also described are probabilistic kernels that can be used for a regression of location, which can achieve up to about inn accuracy in an office environment.

The invention provides a method for recognizing the tool abrasion degree of a large numerical control milling machine. The method includes the steps of obtaining a vibration time domain signals in the running state of the large numerical control milling machine, obtaining frequency domain distribution of the vibration signals through fast Fourier transform, selecting multiple statistical characteristic parameters in the time domain and the frequency domain for conducting dimensionality reduction of tool abrasion characteristic parameters on the basis of a diffusion mapping method, selecting scale parameters of diffusion mapping through the leave-one-out cross validation algorithm and the optimization search algorithm, and recognizing the abrasion degree of an unknown tool to be measured through the combination with Nystrom expansion and the kernel regression algorithm. By means of the method, the disadvantage that a tool abrasion sample of the large numerical control milling machine is missed can be effectively overcome, the recognition accuracy of the tool abrasion degree of the large numerical control milling machine can be improved, and maintenance cost and time caused by untimely tool abrasion recognition can be reduced.

The invention discloses a remote sensing image change detection method based on controllable kernel regression and superpixel segmentation. The problems that only grey information of an image is considered when a difference chart is structured, other feature information is underused, k-means clustering is directly carried out on the difference chart, and therefore the situation that a weak declension area cannot be detected is easily caused are mainly solved. The method comprises the steps of adopting the controllable kernel regression on two input time phase images to respectively extract structural feature matrixes, combining feature matrixes of neighbourhoods with the structural feature matrixes respectively, obtaining a local structural feature matrix, decomposing the local structural feature matrix by using a non-negative matrix factorization algorithm, carrying out a difference chart structure on an obtained coefficient matrix, finally segmenting the difference chart to obtain an over-segmentation image by using a superpixel segmentation method, carrying out the K-means clustering on the over-segmentation image, and obtaining a change detection result. The remote sensing image change detection method based on the controllable kernel regression and the superpixel segmentation can keep marginal information of images, is good in noise proof performance, improves change detection precision, and can be applied to fields of disaster situation monitoring, land utilization, agricultural investigation and the like.

A monitoring system for determining the future behavior of a financial system includes an empirical model to receive reference data that indicates the normal behavior of the system and input pattern arrays. Each input pattern array has a plurality of input vectors, while each input vector represents a time point and has input values representing a plurality of parameters indicating the current condition of the system. The model generates estimate values based on a calculation that uses an input pattern array and the reference data to determine a similarity measure between the input values and reference data. The estimate values, in the form of an estimate matrix, include at least one estimate vector of inferred estimate values, and represents at least one time point that is not represented by the input vectors. The inferred estimate values are used to determine a future behavior of the financial system.

A method for determining the future operational condition of an object includes obtaining reference data that indicates the normal operational state of the object, and obtaining input pattern arrays. Each input pattern array has a plurality of input vectors, while each input vector represents a time point and has input values representing a plurality of parameters indicating the current condition of the object. At least one processor generates estimate values based on a calculation that uses an input pattern array and the reference data to determine a similarity measure between the input values and reference data. The estimate values, in the form of an estimate matrix, include at least one estimate vector of inferred estimate values, and represents at least one time point that is not represented by the input vectors. The inferred estimate values are used to determine a future condition of the object.

A monitoring system for determining the future operational condition of an object includes an empirical model to receive reference data that indicates the normal operational state of the object and input pattern arrays. Each input pattern array has a plurality of input vectors, while each input vector represents a time point and has input values representing a plurality of parameters indicating the current condition of the object. The model generates estimate values based on a calculation that uses an input pattern array and the reference data to determine a similarity measure between the input values and reference data. The estimate values, in the form of an estimate matrix, include at least one estimate vector of inferred estimate values, and represents at least one time point that is not represented by the input vectors. The inferred estimate values are used to determine a future condition of the object.

The invention discloses a multi-frame low-resolution image super-resolution reconstruction method based on a convex combination mode, belongs to the field of computer vision, mainly solves a problem of image processing precision and provides an optimization algorithm. The method comprises the following steps of applying joint motion estimation and super-resolution reconstruction methods to an input multi-frame low-resolution (LR) image sequence to obtain a joint high-resolution (HR) image Ijoint; applying a reconstruction method based on improved kernel regression to the input multi-frame LR image sequence to obtain a kernel regression HR image Ikernel; and applying a convex combination frame provided by the invention and deciding a fused region according to an airspace fusion rule, deciding frequency components that need to be fused according to a frequency domain rule and deciding weight distribution of fusion of two methods according to the global weight, and then fusing the HR images obtained by the two steps through the three aspects to obtain a final reconstructed HR image.

The invention discloses an infrared image background suppression method based on unsupervised kernel regression analysis, belonging to the image processing field. The infrared image background suppression method solves the technical problems that priori knowledge is needed and self-adaptability is poor in the field of infrared image background clutter suppression. Firstly, a sliding window is set to predict background, and Gaussian function is determined to serve as kernel function for unsupervised kernel regression analysis; a background predicting clutter sample is substituted in the function to calculate unsupervised kernel regression equation, and a central test sample (O) is input into the unsupervised kernel regression equation so as to obtain the predicted value of the central test sample (O); then, the central test sample (O) value subtracts the predicted value; the sliding window moves, the above process is repeated until the whole image is processed, and a background suppression result image is output. The invention can effectively improve target detectability and positioning accuracy of an infrared target recognition and tracking system, an infrared image monitoring system, etc.

In training data, a similarity matrix is generated for each of types of data corresponding to different kernels, and graph Laplacians are formed individually from the similarity matrices. An entire graph Laplacian is defined as linear combination of the individual graph Laplacians with coupling constants. Observation variables and latent variables associated therewith are assumed to form normal distributions, and the coupling constants are assumed to form a gamma distribution. Then, on the basis of a variational Bayesian method, a variance of the observation variables and the coupling constants can be figured out with a reasonable computational cost. Once the variance of the observation variables and the coupling constants are figured out, a predictive distribution for any input data can be figured out by means of a Laplace approximation.

The invention discloses a super-resolution reconstruction method of multi-frame images. A steering kernel regression method is introduced separately to a clustering phase of an adaptive sparse learning model and a regularized reconstruction process; the two methods are enabled to complement each other's advantages and disadvantages, forming a reconstruction algorithm of better performance. Motionregistration is not required, and therefore, reconstruction errors due to registration errors are avoided and the quality of a reconstructed image can be significantly improved; the method is applicable to places including any motion mode. Compared with existing learning-based multi-frame super-resolution reconstruction methods, the simple efficient integrated reconstruction model that integratessteering kernel regression and sparse learning is provided which can make use of both global structural self-similar priori constraints and sparse constraints to carry out regressive estimation, so that edge details of an image can be better maintained, and image high-frequency distortion can be relieved; the method is not limited to motion amplitude or application scenes and accordingly is adaptive to complex environments.

The invention discloses an image super-resolution reestablishing method based on a sub pixel displacement model. First, input images are configured into a grid, and an image with pixels distributed evenly is formed; then, the sub pixel displacement model is provided, so that a new image degradation model is established; then, a displacement estimation algorithm based on gradients is designed to estimate sub pixel displacement quantity, and accordingly a displacement kernel function is established and computed; and finally, according to the new image degradation model, the displacement kernel function and a Taylor series extension rule are used for establishing a kernel regression estimation expression, and accordingly a high-resolution image is reestablished. According to the method, two independent links of interpolation resampling and restoring in kernel regression super-resolution reestablishing are integrated, and the complexity of substep modeling and the ill-posedness of deconvolution are avoided. Meanwhile, the method does not limit the number of observed images, the method can be used for single-frame reestablishing and multi-frame reestablishing, and applicability is enhanced.

The invention provides an image super-resolution rebuilding method based on the confidence coefficient kernel regression. According to the method, a confidence coefficient kernel function is constructed, the confidence coefficient kernel regression is carried out on an image through the confidence coefficient kernel function, interpolation in the image super resolution rebuilding is finished, by the adoption of the confidence coefficient kernel regression, not only is a double weighted mechanism of the self-adaption kernel regression inherited, but also distinguishing of pixel reliability is added, the abnormal value can be better restrained, and the image super-resolution rebuilding is achieved. According to the method, the image super-resolution rebuilding is achieved, the robustness is high, and the method can be widely applied to the fields such as monitoring, remote sensing, military, medical science and video entertainment and has the important value.

A method for determining whether to perform vertical handoff between multiple network. The method comprises obtaining a plurality of selection metrics for each network, calculating, for each of the other communication networks, a predicted utility value from at least the corresponding plurality of selection metrics using a variable kernel regression function, obtaining, for the current communication network, a second plurality of selection metrics; calculating a second predicted utility value for the current communication network from at least the corresponding second plurality of selection metrics using a second variable kernel regression function, comparing each of the predicted utility values for each of the plurality of other communication networks with the second predicted utility value and switching to one of the other communication networks having the highest predicted utility value, if the highest predicted utility value is greater than the second predicted utility value.

An atmospheric disturbance image recovering method based on variation regularization relates to image processing. The atmospheric disturbance image recovering method comprises: acquiring a video frame subjected to atmospheric interference of a fixed target or carrying out additional simulated disturbance on one picture by using simulation software to generate one group of video frame; carrying out low-rank decomposition on the video frame to obtain an initial reference image; optimizing the reference image by using an optimization model based on a non-partial total variation regular term and a controllable kernel regression regular term, and accelerating the optimization process by using a separated Bregman algorithm; carrying out B spline interpolation registering on the video frame by using an optimized reference image to obtain a registered video frame; using space weighted nuclear norm minimization to fuse the registered video frame to form a near diffraction limit picture; and carrying out deconvolution processing on the near diffraction limit picture to obtain the picture which is finally deblurred and has no noises. By virtue of the atmospheric disturbance image recovering method, the disturbance removing result is improved and a recovered image with clear vision and abundant details is obtained, so that the atmospheric disturbance image recovering method can be used for observing lands in the air, remotely monitoring and the like.