Research On Image Denoising Algorithm Based On ADMM

In the current society,with the rapid development of science and technology,people’s life is also colorful.In this process,there is an urgent need for digital image processing.Image processing technology is needed in many fields,so it is very important to process the image.In the image transmission,there may be a variety of different effects.For example,the more typical is the impact of noise and transmission time overhead.This will make the image unable to achieve the desired effect.Therefore,image processing and image denoising are worth studying.In this paper,ADMM is studied and applied to the field of image denoising.ADMM is mainly the principle of divide and conquer,which is to divide the original mathematical model into two or more,and then iterative parallel processing.In the learning process,the knowledge of compressed perception and entropy function are combined.Compressed sensing is a hot research topic at present,the main purpose is to reduce the transmission overhead.Among them,the lp-norm of compressed perception can be solved by entropy function,so the entropy function is also taken into account.The main research results of this paper are as follows:Firstly,the denoising effect of greedy reconstruction algorithm is not very good,mainly because the reconstruction accuracy is not high and the reconstruction time is expensive.This paper studies the idea of entropy function which can replace lp-norm and ADMM divide and conquer.Through analysis,we can know that we can combine the two,so we propose an excellent image denoising algorithm based on ADMM.Secondly,aiming at the problem that the image denoising effect is not good,the principle of adjusting entropy function and algorithm framework of pnp-admm are studied.By combining the common greedy algorithm with adjusting entropy function,a new mathematical model is obtained,and then it is brought into pnp-admm framework,thus an image denoising algorithm based on adjusting entropy function AEF-RA-Pn PADMM algorithm is obtained.Finally,the above two algorithms are simulated respectively.The experimental results show that the two algorithms have a certain effect of denoising and reconstruction,and also have a lot of improvement in numerical and time.However,there are still some shortcomings in the proposed algorithm.Next,further research and learning can be carried out in terms of time cost and effect.