Towards a Scalable ESPINA for Neuroscience Data Analysis
ESPINA is an image segmentation tool designed to analyse microscopy images in order to identify neuronal structures and to produce 3D models of these structures. This tool allows to display three-dimensional volumes using auto-stereoscopic monitors. It was initially designed for workstations, but when data volume management or its processing complexity makes unfeasible the implementation of the new tools on these computers, it is necessary to resort to computing servers that delimit response times or by means of scalable solutions and algorithmic optimizations. This paper analyses the migration of this tool from the original implementation to a scalable solution and describes the experience achieved during the development of the workstation version. The proposed alternative is a distributed version of the tool that delegate heavy-computational processes to a cluster, improving the performance of the system in a master/slave architecture.