3D Morphological Tumor Analysis Based on Magnetic Resonance Images

摘要

In this paper, the Glow Cut Algorithm will be applied on MRI data retrieved in order to learn more about tumors morphologically analyze them. In the proposed Glow Cut Algorithm, the tumor area will be analyzed by applying the color threshold criteria. Next, the needed tumor area will be selected, beginning from the outer layer to the inner. The tumor area is analyzed from the area that has the largest amount of pixel size. After dissecting the tumor area, the tumor areas will be calculated based on the four types of morphological coordinates: AT, ATT, ADT, and TN. Using and applying the method described on the brain tumor of a thirty year old human male, the following comparative experiments have been conducted. In order to demonstrate the effectiveness of the proposed scheme, a 3D printer is used to extract the morphology of the tumor. Results have shown that using the four morphological methods and automatic methods have made little difference, showing only an average of a 3% error rate, which enables this paper to suggest a more efficient method than the current manual method.