On automation and certification of a homological method to process biomedical digital images

Abstract

In this paper a methodology to extract and compute homological information from biomedical images is proposed; automating some processes, up to now, manually done. The main features of our approach are the usage of several programming languages (Java, Common Lisp and Haskell) and the application of formal methods (namely theorem provers) to verify the correctness of some of the automated process. As case study to test the suitability of our approach, we have applied it to measure the number of synapses of a neuron.