Kernel adaptive filter

About: Kernel adaptive filter is a research topic. Over the lifetime, 8771 publications have been published within this topic receiving 142711 citations.

Systems and methods for bandlimiting anti-noise in personal audio devices having adaptive noise cancellation

Abstract: A method may include adaptively generating an anti-noise signal from filtering a reference microphone signal with an adaptive filter in conformity with an error microphone signal and the reference microphone signal The method may also include adjusting the response of the adaptive filter by combining injected noise with the reference microphone signal and receiving the injected noise by a copy of the adaptive filter so that the response of the copy is controlled by the adaptive filter adapting to cancel a combination of the ambient audio sounds and the injected noise and controlling the response of the adaptive filter with the coefficients adapted in the copy, whereby the injected noise is not present in the anti-noise signal and wherein each of a sample rate of the copy and a rate of adapting of the adaptive filter is significantly less than a sample rate of the adaptive filter

Fast coupled adaptation for sparse impulse responses using a partial haar transform

Abstract: This work presents a novel scheme for identifying the impulse response of a sparse channel. The scheme consists of two adaptive filters operating sequentially. The first adaptive filter adapts using a partial Haar transform of the input and yields an estimate of the location of the peak of the sparse impulse response. The second adaptive filter is then centered about this estimate. Both filters are short in comparison to the delay uncertainty of the unknown channel. The principle advantage of this scheme is that two short adaptive filters can be used instead of one long adaptive filter, resulting in faster overall convergence and reduced computational complexity and storage. The scheme is analyzed in detail for a least mean squares (LMS) LMS-LMS type of structure, although it can be implemented using any combination of adaptive algorithms. Monte Carlo simulations are shown to be in good agreement with the theoretical model for the behavior of the peak estimating filter as well as for the mean square error (MSE) behavior of the second filter.

Structure Adaptive Anisotropic Filtering for Magnetic Resonance Image Enhancement

Abstract: This paper presents a structure adaptive anisotropic filtering method and its applications. It differs from other techniques in that instead of using local gradient as a means of controlling the anisotropism of the filters, it uses the local intensity orientation of level contours and its curvature to control the shape and the extent of the filter kernel. This ensures that edges and corners are well preserved throughout the filtering process.

An adaptive inverse halftoning algorithm

Abstract: A class of inverse halftoning algorithms that recovers grayscale (continuous-tone) images from halftone images is proposed. The basic structure is an optimized linear filter. Then, a properly designed adaptive postprocessor is employed to enhance the recovered image quality. Finally, a multistage space-varying algorithm is developed that uses the basic linear filter structure as before but with spatially adaptive parameters.

Motion adaptive deinterlacing with accurate motion detection and anti-aliasing interpolation filter

Abstract: Both the motion-detection and infra-field interpolation filter are important factors affect the efficiency of motion adaptive de-interlacing. New accurate motion detection (AMD) algorithm is proposed to improve the accuracy of motion detection, which reduces the possibility of error motion detection with a median filter. To improve the efficiency of intra-field interpolation deinterlacing in moving regions, an anti-aliasing interpolation filter (AAIF) is proposed, which is better than the typical windowed sine function. The simulation results show that peak signal noise ratio (PSNR) of our proposed deinterlacing method is 0.5-7.5 dB higher than that of previous studies and attains the best quality of subjective view.