3. Segmentation and analysis

Segmentation is the process of detecting objects of interest, i.e. isolating them from the background, usually by marking them with labels in order to separate them digitally. The segmentation process depends on the expectations of domain experts. In general, it's a complex task. It can be facilitated by controlled acquisition conditions, such as in machine vision, where lighting, optics and acquisition speed or positioning parameters are set to highlight objects.

From a mathematical point of view, segmentation is a partitioning of the plane into regions, and more generally of the image definition space D. This is the most important step in the processing chain, since good segmentation implies actions or analyses in line with expectations. Subsequently, techniques for characterization and pattern recognition, or even tracking in the case of videos, will be implemented....