Large amount of information is random in image processing. To some extent, type-1 fuzzy sets can reduce the effect of uncertainties. Fuzzy sets theory can provide us with knowledge-based and robust tools for image processing. But there are many situations where it needs to be extended to type-2 fuzzy sets because it can also be difficult to determine the crisp membership function itself. By computing the fuzziness of the pixels' corrupted degree, a new image filter based on type-2 fuzzy sets for impulse noise is presented in this paper. Comparing with some other filters, this new filter is more effective. In the end, simulation results show that the new algorithm is feasible.
展开▼
