Adaptive image filter design

Image processing methods are used to interpret images which are obtained from digital and analog cameras, and medical images which are obtained from medical equipments such as x-ray, computerized tomography, magnetic resonance, nuclear medicine, ultrason etc. Nevertheless, the obtained images are not directly used because of noises and corruptions occurred during acquation. Therefore they need filtering process in order to perform image analysis. The differential equation based image filtering techniques have been used since 1980. However these methods destroy edge in image while denoising. In this study, a novel algorithm was developed to filter gray scale and color images. Similarity measurements of pixels are used as diffusion coefficients. Moreover, three different mathematical approaches for similarity calculations are proposed initially and flowingly fuzzy logic based method was employed. It has been observed in simulations and applications that the edge preserving capabilities of the proposed algorithm are better than conventional approaches during filtering process.