This paper introduced the principle of Diffusion-MRI and Diffusion Tensor Imaging (DTI),and putative relation of D-map. There is difference of size, shape and structure between different human brains, just as there is difference of appearance between different human bodies. To remove the difference, the brain D-maps of seven healthy volunteers were normalized. Then registration and fusion based on wavelet pyramid were implemented to the maps to output parameter images that had statistical meaning,which laid a foundation for computer automatic recognition and computer aided diagnosis. Primary achievement had been made in theory, algorithm and experiment, and experiment result was also given. In future research, what will be needed is only improving precision and SNR.%本文阐述了扩散-MRI,扩散张量成像(DTI)的原理以及平均扩散率D-map的导出关系.由于不同人脑大小,形态,结构跟人体相貌有差异一样也存在大小的差异,本文对七个健康志愿者的脑D-map进行了归一化处理.之后利用基于小波金子塔的图像配准、融合方法把七个人脑D-map配准后融合在一起,导出有统计意义的参数图像.为将来计算机自动认读图像,辅助诊断铺平道路.理论、算法、实验都已取得初步成功并给出了实验结果.只是需要进一步提高精度,提高信噪比.
展开▼
