Ringing is an annoying artifact that is frequently encountered in lossy compressed image formats, intra frame formats and graphics. A novel no-reference metric that can automatically detect ringing annoyance in compressed images is presented. It is a no-reference approach because the quality prediction is based on the distorted image only. The approach consists of two steps: first detecting regions where ringing can occur, and second quantifying the ringing annoyance in these regions. The method considers physical structure of ringing artifacts combined with properties of the human visual system (HVS). For detection of ringing regions a strong edge detector is used to capture the regions with ringing artifacts and an efficient model of visual masking is used to determine the ringing visibility. To quantify the ringing, the visibility of ringing is compared to the activity of local background. The local annoyance score calculated for each individual ringing region is averaged over all ringing regions to yield a ringing annoyance score for the whole image.

1. J. Canny, “A computational approach to edge detection,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 8, no. 6, pp. 679–698, 1986.
2. S. H. Oguz, Y. H. Hu, and T. Nguyen, “Image coding ringing artifact reduction using morphological post-filtering,” in Proc. IEEE 2ndWorkshop on Multimedia Signal Processing.
3. P.Marziliano, F. Dufax, S.Winkler, and T. Ebrahimi, “Perceptual blur and ringing metrics: Application to jpeg2000,” Signal Processing: Image Commun.004.
4. Mei-YinShen and C.C.Jay Kuo,”Review of post processing techniques for compression artifact removal” in Jouranl of Visual Communication and image rpresentation, vol 9,No 1.
5. H. Liu, N. Klomp, and I. Heynderickx, “A no-reference metric for perceived ringing,” in Proc. 4th Int. Workshop on Video processing and Quality Metrics for Consumer Electronics, 2009.

Ringing No-reference metric Texture masking Edge detection JPEG compressed image