Algorithms of three-dimensional filtration using the fuzzy-set theory for color image sequences degraded by noise

A study was conducted, to propose original algorithms of three-dimensional (3D) space-time filtration. The algorithms were based on the theory of fuzzy sets, which represent a new class of nonlinear filters, used for eliminating the effect of additive noise in video images. The proposed algorithms use the value of multi-channel pixels and angular differences, for jointly filtering neighboring images on the basis of new fuzzy-logic rules. These fuzzy-logic rules make it possible, to select pixels of similar structure, significantly increasing the processed sample volume and improving the filtration quality. The criteria that were used in describing and comparing the known and proposed algorithms, were the peak signal-to-noise (PSNR) in decibels, the mean absolute error (MAE), and determining the quality of reconstruction of fine details in the image.