General Adaptive Neighborhood Mathematical Morphology

This paper aims to present a novel framework, entitled General Adaptive Neighborhood Image Processing (GANIP), focusing on the area of adaptive morphology. The usual fixed-shape structuring elements required in Mathematical Morphology (MM) are substituted by adaptive (GAN-based) spatial structuring elements. GANIP and MM results to the so-called General Adaptive Neighborhood Mathematical Morphology (GANMM). Several GANMM-based image filters are defined. They satisfy strong morphological and topological properties such as connectedness. The practical results in the fields of image restoration and image enhancement confirm and highlight the theoretical advantages of the GANMM approach.