A new MAR-PM image filtering algorithm

摘要

In order to solve the over-diffusion problem that often appeared in traditional PM filtering algorithm, a new PM model combined with multi-scale auto regressive (MAR) is proposed in the paper. Firstly, a series of multi-scale images can be got by MAR .Then all these images are filtered by traditional PM model and got a forecast distinct image. With this forecast result to readjust the coefficients of the PM model, the original images can be well filtered finally. The PSNR value is used to judge the efficiency, which shows the efficiency of the proposed new algorithm is much better than the other methods.