This article explains how to use the Compartments operation with imported Atlas objects to get segmentation results per atlas regions.
Contextualising segmentation to Atlas regions
Background
Anatomical atlases are valuable tools in contextualizing segmentation results by allowing users to not only get object counts and morphologies, but also knowing how these change in various regions of an organism or organ. This article does not show how to import and register atlas objects to a sample dataset but only covers how to use pipelines that take imported atlas objects together with image segmentation to contextualise the information.
Analysis Results
The results fo the pipeline are compartmentalised objects. There are several ways to review and interpret the data in the Objects table. Selecting the "Atlas" tag will display all the atlas objects and their features, which could include any morphological features of the objects (volume, surface area, sphericity, etc.), but can also include data concerning the children objects compartmentalized within them, such as the number of children, but also statistical values like numerical density which can be defined as a custom feature by dividing the number of children by the volume of the parent.
The objects table can also be arranged in the so called "Master-Detail" view where parents, track or groups appear in the top part of the table, and the children of these appear in the bottom. In such a layout it is then possible to view not just features of the groups, but also of the objects within them.
image courtesy: Kirsty Craigie, University of Edinburgh, UK.