Alternating Sequential Filters by Adaptive-Neighborhood Structuring Functions

In this work we propose the extension of a well-known class of morphological filters, the Alternating Sequential Filters (ASFs), to include the paradigm of Adaptive-Neighborhood Image Processing, leading to what we have called the Adaptive-Neighborhood Alternating Sequential Filters (ANASFs). By using synthetic and real images to which Gaussian noise was added, we demonstrate the better performance of the open-close and close-open ANASFs against the correspondent ASFs, both from a quantitative and qualitative point of view.