Evaluation of GA Template-matching Method by Simulation Using Chest Helical X-ray CT Images

概要

論文の詳細を見る
Template matching based on genetic algorithm (GATM) we previously developed is a method to detect object(s) existing in a large searching area effectively. In this paper, we investigate the performance of the GATM method by two simulation studies. The first simulalion study was performed to investigate potential effectiveness of the GATM for search and detection of only a detection object put in the observed images. As a result of the first simulation, all objects were correctly detected by the GATM method with about only 10% of searching iterations required by conventional TM. In the second simulation study, by assuming that the GATM was used to detect nodules in chest helical CT images, the GATM was performed to detect multiple objects in an observed image. The performances of the GATM without sharing method, GATM with sharing method and conventional TM were compared. The result showed that the performance of the GATM with sharing was comparable to that of conventional TM, and was better than that of the GATM without sharing. Furthermore, we discussed about the characteristic of our GATM method. From the results of the two simulations, we concluded that the GATM method was effective to detect lung nodules in helical X -ray CT images
Medical Imaging and Information Sciencesの論文