期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2014
卷号：69
期号：2
出版社：Journal of Theoretical and Applied
摘要：An adaptive based artifacts removal algorithm is proposed for removal of blocking artifacts, strip lines, drop lines, blotches and impulses in images. The algorithm detects noise variance level and then proper method is selected depending upon the variance. The algorithm changes the maximum size of window during the filtering operation depending on noise level. The output of the filter is a particular value which replaces the current pixel value at that point on which the value is centered at that time. Thus window size is automatically modified based on the density of noise in the image. It replaces a number of independent algorithms required for removal of blocking and other different artifacts and gives better result. The performance of the algorithm is evaluated in terms of Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), Correlation, Feature Similarity (FSIM), Mean Structural Similarity (MSSIM) and Visual signal-to-noise ratio (VSNR). The computation time is compared with other algorithms that already exist.
