Nonlinear JPEG image filtering at stage of preprocessing
Rostislav I. Sokolov, Renat R. Abdullin · 2016
This paper is devoted to development of nonlinear filter of Markov processes for JPEG images at the stage of preprocessing, allowing to increase the filtration quality when off-line and on-line image recovering. The analysis of wide interference class affecting the image is held. The proper processing algorithm is developed for each kind of noise. Evaluating the synthesized algorithm effectiveness is founded on the image processing digital simulation of the signal mixed with the Johnson and white Gaussian noises. Digital experiment has proved the nonlinear filtering of Markov processes shows an advantage over the median filtering for both fluctuation and pulse noise.