30 results about "Gradient noise" patented technology

A method of cancelling alien noise in coordinated DSL lines, a method of smoothing an alien noise covariance estimate, and a processor and modem for cancelling alien noise in coordinated DSL lines. In one embodiment, the method of cancelling alien noise includes: (1) estimating alien noise vectors for at least some training symbols, (2) arranging the alien noise vectors in a matrix dimensioned for a number of coordinated DSL lines, (3) orthonormally transforming the matrix into a lower-triangular matrix and (4) computing alien noise prediction filters from the lower-triangular matrix.

A system and method for simultaneously reconstructing magnetic resonance images and correcting those imaged for gradient nonlinearity effects are provided. As opposed to conventional methods for gradient nonlinearity correction where distortion is corrected after image reconstruction is performed, the model-based method described here prospectively accounts for the effects of gradient nonlinearity during reconstruction and implements a spatial support constraint to reduce noise amplification effects.

A system's intrinsic mottle & graininess levels are either measured or predicted based on the knowledge of system states including material states and actuator states. The optimal amount of total graininess is then calculated based on established visual models or customer surveys. The difference between this optimal graininess and the machine's intrinsic graininess is then used to calculate the required addition of grainy noise to the print. This additional amount of grainy noise can be added to the print either through textured halftone screens or through noise injection into the input digital (contone) image.

An image processing apparatus receiving a pixel value of a multivalue image includes: an input noise generating section that generates a noise; a green noise generating section that generates a green noise from an output value of a binarized processed pixel; an adder section that adds an error integrated value, the green noise and the noise, to the pixel value of the target pixel; the threshold value processing section that binarizes the pixel value of the target pixel after adding the error integrated value, the green noise and the noise; a subtractor section that calculates an error value by calculating a difference between an output value of the binarized target pixel and the pixel value of the target pixel including the error integrated value and the noise; and an error integrating section that outputs the error integrated value by using the error value of the binarized processed pixel.

The invention belongs to the technical field of neural information, relates to a method for extracting EEG signals, and specifically provides an extraction method for EEG signals in a magnetic resonance environment, which is used to improve the efficiency and accuracy of extracting weak EEG signals while retaining wide-band EEG signals. Electrical information, so that it can be widely used in clinical and scientific research; the invention first records the EEG signals without magnetic resonance interference, and then records the EEG signals in the magnetic resonance environment, and then uses two-stage adaptive noise cancellation to remove the magnetic resonance signals in turn. Gradient noise and power frequency interference, and finally a smoothing filter is used to remove high-frequency noise to obtain high-quality EEG signals. The invention can effectively remove magnetic resonance gradient noise, power frequency interference, and high-frequency noise of instruments, extract high-quality EEG signals, and greatly improve signal extraction efficiency and precision.

The invention discloses a document image binarization method based on the symmetry of stroke structure. Wherein, the method includes: determining the gradient image of the document image, wherein the document image is a grayscale image; using the maximum inter-class variance method to perform global binarization processing on the gradient image; according to the width and The symmetry of the gradient direction in the local area, remove the non-stroke gradient noise in the image after global binarization processing, and determine the gradient image with symmetric local gradient direction; determine the structural symmetry element image based on the gradient image with symmetric local gradient direction; The local density of foreground elements in the structurally symmetric element image is used to filter out noise, and combined with the document image to perform local binarization based on a voting strategy. The embodiment of the invention solves the technical problem of how to enhance the adaptability to document image text extraction.

The invention discloses a brain signal online denoising method in a nuclear magnetic resonance environment. The method comprises the steps of 1) conducting high-pass filtering on collected brain electrical signals; 2) conducting up-sampling on the filtered signals and synchronizing the brain electrical signals with markers sent by a fMRI device; 3) constructing a sliding window to construct a noise template Atau for preliminary denoising; 4) slicing a signal Sh mixed with noise into a slice with a length of T*w points, fitting the noise template with slice date through least squares to obtainfitting parameters ytau, and subtracting ytau*Atau from the Sh to obtain a signal Sr with residual artifacts; 5) conducting PCA on the signal Sr, sorting each component according to degree of relevance, taking the largest m components as an optimal base betaj(j=1,2,...,m) of gradient noise, fitting the optimal base with the Sr through the least squares to obtain fitting parameter aj(j=1,2,...,m),subtracting N from the Sr, and conducting denoising through the construction of a gradient noise template and denoising through the PCA to eventually eliminate the gradient noise. The brain signal online denoising method in the nuclear magnetic resonance environment has the advantages that real-time denoising can be conducted on the signals to meet demands of data processing speed of an online experiment.