International Journal of Computer Applications |
Foundation of Computer Science (FCS), NY, USA |
Volume 65 - Number 5 |
Year of Publication: 2013 |
Authors: B. R. Benujah, X. Jushwanth Xavier, S. S.uma |
10.5120/10918-5851 |
B. R. Benujah, X. Jushwanth Xavier, S. S.uma . Frame Work for Auto Segmentation of Fibrocartilaginous Disc of Scoliosis Affected Spine. International Journal of Computer Applications. 65, 5 ( March 2013), 7-11. DOI=10.5120/10918-5851
A framework for automatic segmentation of fibrocartilaginous disc of scoliosis affected spine from MRI image is presented in this paper. This method uses a combination of statistical and spectral texture features for discriminating closed regions representing fibrocartilaginous disc from background in MR images of the spine. Texture features are extracted from the closed regions based on the watershed approach. The feature selection step is based on principal component analysis and clustering process. It permits to decide among all the extracted features which ones resulted in the highest rate of good classification. With the help of the selected texture features and classification, the problem of over-segmentation underlying in existing automatic segmentation methods can be solved successfully by discriminating fibrocartilaginous disc from the background on MRI of scoliotic spines.