Segmentation of Brain Tumor from MRI Image using Gradient Vector Flow

Brain tumor scans are one of the complex segmentation problems. It is not at all easy to extract the tumor affected regions from the surrounding pixels which makes it a really diffi cult aspect for computer aided diagnosis systems. A new vector fi eld analysis of the GVF images for the detection of tumor boundaries has been proposed by the author. The boundary and non-boundary points has been identifi ed using the vector entropy analysis algorithm of GVF vectors. The output of the proposed algorithm has been compared with other existing segmentation methods on the basis of three parameters-sensitivity, specifi city and accuracy. Experimental result shows that our proposed scheme outperforms previous schemes based on other approach.