I'm trying to implement the Adaptive Morphological Edge-Linking Algorithm (Frank Y. Shih, Image Processing and Pattern Recognition: Fundamentals and Techniques, section 5.5.2 OR Article link). The problem is I'm having trouble understanding how to implement the adaptive morphological dilation and how it should be applied (from the algorithm's description, I understand that I have to apply adaptive dilation at each end point, but wouldn't this affect the rest of the image too? I think I'm missing/misunderstanding something)
EDIT: Actually, I think I understand what it means by applying at each end point. Apparently, each point in the image has its own kernel or so. It does seem to say something about this in the operation's definition. I'll look more into it, hopefully I can find a way to implement it.