SEGMENTATION OF UTERINE USING NEIGHBORHOOD INFORMATION AFFECTED POSSIBILISTIC FCM AND GAUSSIAN MIXTURE MODEL IN UTERINE FIBROID PATIENTS MRI

Abstract

Uterine fibroids are common tumors of female pelvis. Uterine volume measurement before and after surgery has an important role in predicting the outcome and later on in comparing with the result of the uterine fibroid shrinkage surgery. Because of inhomogeneity and different shapes and sizes of uterus and fibroids, segmentation of uterus is a difficult task. In this paper, using T1 and Enhanced-T1 MR images uterine is initially segmented using a new clustering algorithm named neighborhood information affected possibilistic fuzzy C-means (NIAPFCM). NIAPFCM uses membership, typicality and spatial neighborhood information to cluster each voxel. Finally, the redundant parts are removed by superimposing the segmented region of the T1-enhanced image over the registered T1 image. Gaussian mixture model (GMM) is applied to the extracted region histogram as a model for accurate tresholding. The results obtained using the proposed method are evaluated by comparing with manual segmentations using volume-based and distance-based metric methods. Also, the result of NIAPFCM is compared with fuzzy C-means (FCM) and possibilistic fuzzy C-means (PFCM) algorithms. We found this algorithm efficient, which provides good and reliable results.