Segmentation of CT Head Images (CROSBI ID 465697)
Prilog sa skupa u zborniku | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Lončarić, Sven ; Ćosić, Dubravko ; Dhawan, Atam P.
engleski
Segmentation of CT Head Images
Segmentation of human head images obtained by computed tomography (CT) plays a central role in intelligent image analysis. A new method for CT head image segmentation is presented in this work. In particular, segmentation of human spontaneous intracerebral brain hemorrhage (ICH) is important for quantitative analysis of ICH The proposed procedure classifies each CT image pixel into one of the following regions: background, skull, brain, ICH , and edema. CT head image segmentation has shown to be a challenging task. Most regions are relatively well localized but there is a lot of ambiguity in the edema localization. The proposed method consists of two main phases. An unsupervised fuzzy clustering algorithm is used in the first phase to generate a number of spatially localized image regions having uniform brightness. The unsupervised fuzzy clustering algorithm used in this work is a combination of the fuzzy C-means algorithm and the fuzzy maximum likelihood estimation. The unsupervised algorithm is used because no prior knowledge about the number of clusters is available. In the second phase an image labeling algorithm is used to merge multiple clusters of similar properties into unique image regions. The backtracking tree search algorithm has been used to find solutions of the labeling problem. The algorithm assigns a label to each of the small regions resulting from the clustering phase. The label set consists of five elements: background, skull, brain tissue, hemorrhage, and edema. Constraints are imposed on the solution of the labeling algorithm using region neighborhood relations and region label relations. The proposed method has been tested on real CT head images and has shown satisfactory results.
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
1012-1012-x.
1996.
nije evidentirano
objavljeno
Podaci o matičnoj publikaciji
Proceedings of the International Symposium on Computer and Communication Systems for Image Guided Diagnosis and Therapy
Lemke, Heinz U. et all.
Elsevier
Podaci o skupu
International Symposium on Computer and Communication Systems for Image Guided Diagnosis and Therapy
poster
01.01.1996-01.01.1996
Pariz, Francuska