Research On Image Denoising Method Based On NSST Domain Statistical Model

With the rapid development of science and technology,as the carrier of multimedia communication,the application of image in life becomes more and more extensive.Due to the external or insitu imaging system,the image produces a lot of noise during the transmission.These noise not only affect the visual effects of the image,but even change the content and quality of the image,which cause great interference in image segmentation,image retrieval,feature extraction and other subsequent digital image processing operations.Image denoising technology is to solve that problem.The purpose of image denoising is to filter out the noise from the noise polluted image and get the original "pure" image,but it also need to ensure that the image of the internal edge and texture structure are not affected too much.When denoising in the frequency domain space,it can be found that the larger the coefficient value is,the more image information is contained.By counting its own statistical properties,more accurate reduction of the original coefficient values at the noise point can be got.Based on this theory,three image denoising algorithms based on non-subsampled shearlet transform(NSST)are proposed.1.An Image Denoising Algorithm of NSST Domain Based on priori Cauchy Model.The Cauchy distribution,as a prior distribution model,is used to fit the probability distribution of the subband coefficients of the NSST transform domain,and the noise-free coefficients are then estimated by the maximum a posteriori probability(MAP)method.This method not only preserves the advantages of the traditional statistical model denoising method,but also uses the Cauchy distribution model with better fitting effect of the subband coefficients of NSST as the prior probability distribution model,so that the estimated coefficient is closer to the coefficient of the original image.2.An Image Denoising Method Based on Cauchy Distribution NSST-HMT Model.Through the joint distribution of NSST subbands,we obtain the persistence and aggregation of NSST parent-child coefficients.Combining the probability distribution of NSST coefficients conforms to the Cauchy distribution,the C-NSST-HMT model is proposed and applied to image denoising.This method not only considers the correlation of coefficients in subband,but also considers the correlation between parent and child coefficients.The experimental results show that the method can not only obtain the estimation of noise-free coefficient,but also can protect the edge of the image content.3.NSST-HMT Model Based on Joint Distribution Statistics of Parent-Son Coefficients and Brother Coefficients and Its Denoising Application.Through the joint statistics betweenthe NSST brothers subband coefficients,it can be found that the NSST brother coefficients have similar statistical characteristics with parent-son coefficients,Regarding which as a reference factor for HMT model state transition,the C-F&B-NSST-HMT model is proposed and applied to image denoising.This method not only considers the correlation between subbands in the same direction on different scales,but also considers the subband correlation in different directions of the same scale.The experimental results show that the proposed method is more accurate and has stronger edge protection ability.NSST is the most powerful multiscale tool to capture the important features of two-dimensional images and Cauchy distribution has the strong ability to describe the NSST coefficient distribution with the characteristics of "peak heavy tail".Therefore,the NSST coefficients are modeled by Cauchy distribution.On the basis of that,the latter two algorithms of this paper fully consider the correlation between NSST coefficients and protects the edge and texture information of the image while realizing image denoising.