Supporting the recognition of pathological changes in CT coronary arteries visualizations based on data aggregation approach

In this paper, the data aggregation approach is proposed to support the computer-aided detection of pathological changes in coronary arteries structure. Unlike the previous papers the authors mainly focused on the possibilities of utilization of data aggregation techniques to assess the degree of threat to the patient's health depending on the type of distribution of coronary arteries (balanced artery distribution, right dominant artery distribution or left dominant artery distribution). Moreover, in order to determine the appropriate projection for further analysis from the set of image data the authors proposed two heuristics: based on the Feret diameter and based on the center of gravity. Presented techniques can support a comprehensive description of the analyzed image by identifying the most important parameters of a given structure, and then analyzing their meaning by aggregating information obtained from two sources: (1) the topology of coronary arteries (their mutual spatial relations), and (2) the analysis of individual arteries of the coronary arteries system. The obtained results are very promising - for the image data included in the experiment around 82% of recognitions were correct.