Research On Digital Image Inpainting Algorithm Based On Generalized Gaussian Mixture Model

Image restoration is an important research problem in the field of digital image pro-cessing,which includes digital image denoising,deblurring,inpainting and super resolu-tion.In recent years,as an image restoration algorithm,Expected Patch Log-Likelihood(EPLL)algorithm has attracted wide attention in the academic field.EPLL algorithm is based on statistical prior model and maximum likelihood estimation to process image patch.The advantage of EPLL algorithm is that it can be combined with any statistical prior model,and it can provide excellent image restoration effect in practice.generalized Gaussian mixture model is a kind of statistical prior model which takes Gaussian mixture model and Laplacian mixture model as special cases.When the parameters of the gen-eralized Gaussian mixture model take different values,the generalized Gaussian mixture model can correspond to different mixing models,so the generalized Gaussian mixture model is more flexible.In this paper,the generalized Gaussian mixture model is combined with EPLL algo-rithm,and the whole processing structure of the model and the corresponding algorithm are described comprehensively and systematically.In this paper,based on this algorithm model,the application of the model is extended,and the model is successfully applied to the problems of image inpainting and extension resulting in image super resolution.At the same time,the corresponding numerical solution method is given.In order to illustrate the effectiveness of the results presented in this paper,a large number of appli-cation experiments are carried out on different images using the algorithm presented in this paper,and the experimental results are shown.The experimental results show that the method presented in this paper is feasible and can provide reference for relevant practical applications.