ADVANTAGES OF USING SIFT FOR BRAIN TUMOR DETECTION

Abstract

The brain is the anterior most part of the central nervous system. The cranium, a bony box in the skull protects it. Virtually every activity or thought of ours is controlled by our brain. So, it’s very dangerous when the proper functioning of the brain is hindered. Brain tumor is one such disease which if not detected early and treated accordingly, can prove fatal. Structure of the brain is quite complex and hence it is very difficult to detect the abnormalities in early stages. In our paper we will be giving an overview of the various techniques used for brain tumor detection and how SIFT overcomes their limitations. The techniques discussed include biopsy, manual segmentation, mathematical morphology & wavelet transform, artificial neural network and finally SIFT (Scale Invariant Feature Transform). Biopsy is a surgical method which needs to be performed by highly skilled professionals. The rest other methods use MRI images and thus are non-invasive. SIFT technique which we are using in our project gives good accuracy, is cost effective and most importantly is invariant to translation, scale, rotation, affine transform, change in illumination, etc.