Discrete Mereotopology in Histological Imaging

Gabriel Landini, David Randell, and Antony Galton

In Ela Claridge, Andrew D. Palmer and William T. E. Pitkeathly (editors), Medical Image Understanding and Analysis, Proceedings of the 17th Conference on Medical Image Understanding and Analysis, 17th-19th July 2013, Birmingham, UK, pages 101-106.
ISBN: 1-901725-48-0

Available at http://events.cs.bham.ac.uk/miua2013/MIUAproceedings.pdf

Abstract

In this paper we describe methods suited for developing intelligent histological imaging procedures based on mathematical morphology and a discrete version of the Region Connection Calculus (RCC) known as Discrete Mereotopology. The implementation of the discrete versions of RCC5 and RCC8 relation sets enables computation of the spatial relationships between image regions and reasoning about those relations in segmented digitised images. It also opens the possibility of defining histologically relevant models of biological structures (cells and tissues) so the relations of their components can be assessed algorithmically. A Java plugin implementing the RCC5D and RCC8D relations sets for the popular imaging tool ImageJ was developed. We illustrate an application for automated cell sorting on cultured fibroblasts.


Antony Galton
Last modified: Thu Aug 1 09:59:50 BST 2013