Lung tumor area recognition in CT images based on Gustafson-Kessel clustering
Parnian Afshar, Abbas Ahmadi, M.H. Fazel Zarandi · 2016
Recently, lung cancer has attracted a great deal of interest due to its commonness and deathly nature. By the development of technology, medical imaging methods play important role in both diagnosis and treatment of this cancer. Lung tumor segmentation in CT images is of high importance in many areas like cancer treatment. One of the common methods for segmenting images, used in this article, is image clustering. K-means and FCM clustering which have been widely used in this area have the problem of producing clusters with the same shapes. To overcome this problem, Gustafson-Kessel clustering is adopted in this article. To improve the model's performance, lung is first separated from the background using Snake optimization method. Comparison between our proposed method and two other clustering methods, fuzzy C-means clustering and K-means clustering, using entropy analysis show that our proposed approach is much more effective for the stated purpose.