Iterative noise filtering

Two methods for additive noise filtering are proposed in this thesis. The first method, Automated Noise Filtering (ANF), is a modified and automated version of an already existing noise removal method called Noise Filtering by use of Local Statistics (NFLS). ANF is capable of processing images with different degradation levels. Furthermore, it is completely automated, and thus, requires no a priori knowledge of signal or noise parameters. ANF is able to preserve vital signal information, i.e., edges and details. It is computationally efficient and well suited for parallel processing. The second method. Iterative Noise Filtering (INF), is an iterative version of ANF. INF possesses all the abilities of Automated Noise Filtering, and in addition, it is capable of removing the additive noise completely from the highly degraded images. This iterative process terminates automatically after the additive noise is entirely removed. Performance of the methods of Automated and Iterative Noise Filtering are evaluated based on their applications to various images. Each of the two methods are shown to compare favorably to an alternative technique of noise removal.