Image Denoising Method Using the Gradient Matching Pursuit

Abstract

Image denoising method based on sparse decomposition means the useful information in the image is taken as the sparse component and noises in the image as the residuals after the removal of sparse component, which is the basis for image denoising. In this paper, a new image denoising algorithm using the gradient matching pursuit is proposed based on the study of image sparse decomposition. It firstly constructs an over complete atomic library in the image, then the optimal atom is found by the sparse decomposition method with gradient tracking. At last reconstruct the image with using the optimal atomic. The simulation results show that the convergence of the improved algorithm is fast and stable, and it can effectively eliminate the noise.

Imaging Experiment of Brain Cognitive Activity Based on EEG and Its Philosophical Influence

“…There are two reasons for this: first, from the point of view of the problem concerned in the experiment, Experiment 1 mainly studies the differences of EEG in the process of judging and recognizing sentences by the subjects; Experiment 2 mainly studies the cases of correctly recognizing sentences by the subjects, and the problems that these two experiments focus on are relatively small and not universal; the problems studied in Experiment 3 are the differences between things and concepts of the main problems in the history of philosophy, which is firstly expressed in Berkeley and secondly in Husserl's phenomenology: Is there any difference between the two behaviors when a person sees an object and hears the name of the same of the same object? Secondly, from the experimental method, Experiment 3 is to study the above-mentioned problems by EEG contrast, that's, drawing a conclusion by comparing the EEG images produced by the two psychological activities (Alexiou et al, 2009;Tang and Chen, 2016).…”

An Electromagnetic Detection Method for Grain Silos Based on Finite Difference Time Domain and Ground Penetration Radar

“…Since the silo has a fixed thickness, the n value was selected through trial and error under the constraint of the range of dielectric constraint of the grain [10]. Figure 4(b) provides the results after adding the phase to the periodic factor (2πn, n=1).…”

“…The parameters of the GPR system were configured as: quality factor=90dB, antenna gain=3dB, antenna efficiency: 1/3, and power reflection coefficient of antenna surface=1/3. The electromagnetic parameters of the grain were directly selected from Section 3.2: the real part of the dielectric constant was 4 and the attenuation coefficient was 6dB/m [10].…”

Effect of suction nozzle structure on reverse circulation performance of down-the-hole hammer drill bit

“…The most proper model should be selected against the flow physics, accuracy requirement, computing resources and simulation time. Since the RC DTH hammer drill bit was design with multiple ejectors, it is worthwhile to refer to the previous CFD simulations of ejector flow fields [18][19][20][21][22][23] before selecting the turbulence model. After long deliberation, the author selected the realisable k-ε model for the simulation.…”