Image Segmentation, Fusion and Quantification of Brain Tumors and Nerve Fibers

• Main Authors: Hua-Jiun Chiou, 邱華俊
• Other Authors: Jia-Chi Fu
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/34948448653229689509

碩士 === 雲林科技大學 === 工業工程與管理研究所碩士班 === 99 === Recently, diffusion tensor imaging (Diffusion Tensor Imaging, DTI) and Magnetic Resonance Imaging (Magnetic Resonance Imaging, MRI) is widely used in clinical medicine not only for the delineation of white matter fiber tracts, but also to assess the histological features of the lesions, the relationship between tumor and surrounding tissue excision for cancer, treatment, assessment of the degree of erosion are very important. T1 magnetic resonance imaging and T2 magnetic resonance images can allow the physician to fully understand the tumor and peritumoral edema space information between each other. This thesis brain diffusion weighted imaging (Diffusion Weight Imaging) imaging operations using a diffusion tensor imaging (Diffusion Tensor Imaging) and display images nerve fibers, and with magnetic resonance imaging (MRI) brain tumor and some surrounding edema by manually segmentation the results of image information fusion of information, and finally using quantitative analysis of index: Fractional Anisotropy(FA), Mean Diffusivity(MD) measured four metastatic tumors, edema region, edema symmetric region and four normal subjects the relationship between FA and MD; Tumor Infiltration Index (TII) quantitative analysis and assessment of tumor tissue surrounding nerve fibers by tumor cells eroded some of the damage caused by the impact of invasion. In this thesis, Taichung Veterans General Hospital Radiology Department offers four patients with metastatic tumors and four normal subjects, using diffusion weighted imaging (DWI) in brain tissue provides the main diffusion direction of water molecules, imaging compute the DTI image will be tumor, edema, and DTI imaging regulation and fusion of three-dimensional images and three-dimensional visual output, the image information through the DTI calculation FA and MD values of tumor, edema, and edema of symmetric regions, to evaluate FA and MD of information distributed of tumor, edema, and edema of the symmetrical regions, finally, to analyze the average normal subjects and the correlation between metastatic tumor. Test four metastatic tumor and normal subjects, preliminary results showed quantitative index of FA and MD with considerable stability, because the region of tumor and edema FA and MD of the mean square error(Mean Square Error, MSE) exist above and below the regression line information dissemination point, the linear regression line and can not be correctly classified four groups of metastatic tumor and normal subjects the average FA and MD distribution point of information , and using K-mean classification of the average point, The result certainty fullness divided into two cluster, provide quantitative basis for medical imaging