EFFiCIENCY OF images FILTERING WITH LIMITED APRIORY INFORMATION ON NOISE PROPERTIES

D.V. Fevralev, V.V. Abramova, S.K. Abramov, V.V. Lukin

The filtering efficiency of images corrupted by the signal-dependent noise with limited information about noise type and its statistical properties has been analyzed. The study is carried out for filters based on discrete cosine transform because such filters are able to take into consideration a-priory or retrieved from the images information. It has been shown that the use of a-priori or retrieved from data information in filtering allows increasing filtering efficiency by 0.3…1 dB according to PSNR and PSNR-HVS-M metrics, but it is desirable to improve the evaluation accuracy of signal-dependent noise characteristics.

Key words: DCT-filtering, signal-dependent noise, filtering efficiency, visual quality.