Comparison of various fuzzy clustering algorithms in the detection of ROI in lung CT and a modified kernelized-spatial fuzzy c-means algorithm

The detection of pulmonary nodules in radiological images or Computed Tomography has been widely researched in the field of medical image analysis, because it is a highly complicated but socially interesting matter. The classical approach consists in the development of a CAD system that indicates in phases the presence or absence of nodules. One of these phases is the detection of regions of interest that may be nodules, with the aim of reducing the problem area. This article evaluates various fuzzy clustering algorithms that represent current tendencies in the field, and proposes a new algorithm. The algorithms were evaluated with high resolution CTs from the Lung Internet Database Consortium.