In both two-dimensional and three-dimensional image processing, a basic task is the segmentation of the image into sub-images that each represent a single item of interest. For example, several people standing in a group, or some groceries on the shelf. This thesis topic permits two possible directions of research: first, given (three-dimensional) imaging from a set, distinguish the separate objects in that image (also considering possible partial overlaps) and produce a (two-dimensional) image of each segmented item; second, given a (two-dimensional) image of an object, detect the parts of the that object. In the particular case of products sold in stores, the elements to detect include the logo of the company/product, a barcode or a QR code when present, and the text regions in the product packaging. The ultimate goal is to then send these image segments, separated from each other, to other software modules that interpret them, such as barcode decoders, optical character recognition, and such – the interpretation is out of the scope of this work, as the segmentation itself (of items from a scene or elements from an item) is challenging enough for thesis work as it is.