Mathematical morphology filters using basic structuring elements (erosions, dilations etc) are classics. Most textbooks covering morphological filtering restrict their exposition to these basic filters. This is unfortunate, since very significant advances have been made since the development of these filters, in particular by making the filters adaptive to the image content. This has lead to a whole new range of image processing and analysis tools, ranging from shape preserving image denoising techniques, to object recognition methods. In this tutorial we will discuss the theory behind recently developed advanced morphological filters, including adaptive structural filters, viscous filters, morphological PDEs, path openings, amoeba filters, and connected, semi-connected and hyperconnected filters. Numerous applications will be used to demonstrate the usefulness of these filters. This tutorial is the second in a series, the first having been presented at ICPR 2014 in Stockholm. The present proposal is updated with new results and applications that have recently appeared.
Advanced morphological filters, such as (hyper)connected filters, path openings, and others are not that well known, and at the same time can be adapted to a vast range of image analysis tasks, it is important to bring these filters to the attention of a wider audience. The most important advantage of these filters is that they allow researchers to approach their problem from a radically different direction than more traditional means, such as linear filtering, PDE-based methods, wavelets, or morphological filtering based on fixed structuring elements. As such, these filters do not so much replace, as augment these traditional approaches.