Imaging plays a central role in the diagnosis and treatment <br />planning of brain tumor. An accurate segmentation is <br />critical, especially when the tumor morphological changes <br />remain subtle, irregular and difficult to assess by clinical <br />examination. Traditionally, segmentation is performed <br />manually in clinical environment that is operator dependent <br />and very tedious and time consuming labor intensive work. <br />However, automated tumor segmentation in MRI brain <br />tumor poses many challenges with regard to characteristics <br />of an image. A comparison of three different semiautomated methods, viz., modified gradient magnitude <br />region growing technique (MGRRGT) , level set and a <br />marker controlled watershed method is undertaken here for <br />evaluating their relative performance in the segmentation of <br />tumor. A study on 9 samples using MGRRGT reveals that <br />all the errors are within 6 to 23% in comparison to other <br />two methods
