Automatic detection, extraction and mapping of Brain Tumor from MRI scanned images using Frequency Emphasis Homomorphic and cascaded hybrid Filtering techniques

Magnetic Resonance Imaging is one of the best technologies currently being used for detecting Brain Tumor at both early and advanced stages. This paper proposes a very simple, efficient and automatic technique to detect brain tumor from MRI scanned images. Usually tumor contains more water as well as more Hydrogen atoms than normal tissue. So the intensity of tumor regions in MRI images will be different from normal tissues. This methodology presented here utilizes the intensity difference immaculately by using Frequency Emphasis in Homomorphic Filtering. Especially tumor regions are intensified and noises are reduced by Gaussian equation based algorithm. Hybrid filter is used in cascaded method with spatial domain filters to eliminate time domain noises. Then by thresholding, segmentation and morphological operations the tumor regions are extracted and mapped in edge detected image of brain. This not only provides the size of the tumor region but also its accurate position in Brain.