Noise Detection and Reduction Using Algorithms Based on Particle Swarm Optimization and Fuzzy Decision Making

Noise filtering has a special importance for systems related to digital images processing. Mostly conventional filter used in image processing applications for noise cancellation is the median filter which is especially used to reduce salt and pepper noise in images. However median filter has a good performance in smoothing and noise reduction, but leads to blurring and elimination of some fine details and destruction of edges in image. One approach in confronting with conventional noise filters like median filter is not to operate same process on all pixels. This means by using a decision block or classifier, confronting with pixels must be determined appropriately. In this thesis we address to the problem of designing such decision blocks for noise reduction based on fuzzy inference theory. Another main subject in this thesis is the procedure of utilization of optimization algorithms based on swarm movement of particles for designing the fuzzy part.