Brain tumor segmentation and its study are tricky assignments of medical image processing due to complexity and variance of tumors however, forms a decisive factor for quantitative exploration of the spatial data in magnetic resonance imaging of human brain. In that mode, this modality
of image has developed into a valuable investigative means in medicinal domain for detecting irregularity and discrepancy in human brain. The accuracy of segmentation method relies on its capability to discriminate different tissue, classes, discretely. Consequently there is an essential need
to evaluate this capability prior to employing the segmentation method on medical images. In this paper, a semi-automatic segmentation technique is proposed to carry out the analysis and study of proficient pathologies of brain tumor of human brain. The task of segmentation is carried out
integrating region growing with active contour methodologies. The evaluation of proposed methodology has been carried out on multislice image of MRI data and compared with other semi automatic and automatic techniques. It is observed by the experimental results that proposed system has the
ability to accomplish fast segmentation and exact modeling of tumors in brain with a gratifying accuracy in order to support future treatment planning.
Automatic Segmentation of the Left Atrium From MR Images via Variational Region Growing With a Moments-Based Shape Prior
