Segmentation of brain tumour regions with new active contour

Abstract

Magnetic Resonance Imaging (MRI) is used to get a detailed image of the organs of the body. MRI has a significant impact on diagnosis, growth rate prediction, and in the treatment planning of brain tumour due to its high soft-tissue contrast property. Surgery is the most common treatment for brain tumours. Radiation and chemotherapy are used to slow down the growth rate of a tumour. In all treatment methods, it is crucial to know the size and position of a tumour. Segmentation of brain tumours from MRI images can be used to get these details. However, this is challenging due to inhomogeneity of MR Images and a large volume of data available from 3D scanning modalities. There is no unique or best method for segmentation of MR images. Different MRI modalities are used to segment-specific regions of a tumour. The presence of noise and inhomogeneity in MR Images limits the accuracy of segmentation algorithms. The proposed work presents two new active contour models for segmentation of tumour regions. The first model is a Fuzzy, integrated active contour model; it is a combination of fuzzy and level set methods. The second model is a Modified active contour model, uses a shift in statistical mean values of image regions. These proposed active contour methods are used to visualise tumour in 3D. newline