The first part [1] of this two-part work covers the theoretical aspects of the so-called general adaptive neighborhood image processing (GANIP) approach. The adaptive neighborhood paradigm speaks about image transformation using local image features, based on context dependencies. In the first paper, the morphological operators’ structuring elements are replaced by GAN-based structuring elements. The operators derived in this way are fitted to the local contextual details of the considered image. This second paper addresses the practical aspects of GANIP. To understand the second part fully, I suggest you read the first part first.
The main contribution of this paper is to present several examples of practical applications of the GANIP approach, for example, image filtering, image segmentation, and image enhancement. The effects of these applications are easily observed in the figures presented in the paper.
The use of the GANIP approach in the context of mathematical morphology is an interesting aspect of this paper. After substituting the structuring elements required for mathematical morphology operators with GAN-based structuring elements, we can construct operators by reserving topological properties of the objects.