Certain investigations on brain tumor segmentation using morphometric and back propagation neural networks

Abstract

Tumor differs in shape, dimension and position which gains the properties of typical tissue and overlaps with the abnormal tissues. The symptoms of tumor in human brain largely embrace abnormal changes in ability, vision, hearing problem and seizures etc. To identify and analyze the abnormalities of tumor, various medical imaging techniques like MRI, NMRI, CT, MRT and PET are facilitated. Typically, the use of MRI technique provides the three dimensional information of human brain and used exactly in brain disorder detection, designation and classification. This technique is largely utilized in hospitals to diagnose and detect the stages of tumor without any exposure to radiation. The researcher proposes to detect the tumor in human brain by using conventional segmentation, Morphometric approach and back propagation neural network techniques. Segmentation is the method of dividing a picture into regions with similar properties likes grey level, color, texture, brightness, and distinction. The outcome of image segmentation may be a set of area that conjointly connects the whole image. Every pixel during a region is related with some uniqueness like color, intensity or texture. The segmentation algorithms are the key parts of machine-controlled imaging diagnostic systems. newline