Patterns of image analysis: context dependent bridging of gaps using the example of tube formation assay

Abstract

Consulta en la Biblioteca ETSI Industriales (8875)

[EN] The growth of blood vessels from existing vasculature occurs in both health and disease. The analysis of this phenomenon, called angiogenesis, has grown during the last years. To consider this phenomenon, images must be taken by a microscope, and experts (doctors, scientists) base the results of their analysis (by means of the human eye) in their experience. These images can also be analyzed by computers, however, it is not an easy task. There have been many attempts to try to solve these cases by automated image analysis. Although the capabilities and advances in this way have grown, the automated image analysis has been limited. Human beings consider things in the images as objects and understand the relationships and the context between them. However, computers analyze the pictures pixel-by-pixel, and make the task in fact difficult. The main task in this work is to analyze an image within the tube formation assay, the main software used belongs to Definiens Cognition Network Technology¿, and it is called Definiens Developer XD. By this software it is possible to create objects, consider their relationships and classify them, due to the object-oriented and context-based logic it follows. The aim of the work is finally to join the tubes growing to make the network of vessels, and Definiens Developer XD is not good in image modification. To complete this task, other software, the ImageJ module, has been used, in order to join the tubes. This software uses the pixel-by-pixel identification to make image modification. The automated image analysis has the aim of speed up the process of analyze biomedical images. It is possible to process thousands of these images in a fraction of the time it would take experts.