Data mining based noise diagnosis and fuzzy filter design for image processing

In image processing, both diagnosis of noise types and filter design are critical. Conventional filtering techniques for image restoration such as median filter and mean filter are not effective in many cases, such as the case lacking the information of noise types or the case having mixed noise in images. This paper develops a data mining approach for noise type diagnosis, and proposes a fuzzy filter design for enhancing the quality of noise corrupted images. The experimental results demonstrate that the proposed technique outperforms the conventional filters, particularly for dealing with the images corrupted by mixed noise with additive Gaussian noise and impulse noise. (C) 2014 Elsevier Ltd. All rights reserved.