简介: |
报告摘要：
This talk will summarize our work on analysis of MR brain images. Our main goal is to develop automated analysis methods for precise quantification of subtle and complex structural/functional changes in the brains, with applications in early detection of brain disorders, such as Alzheimer's Disease (AD). To achieve this goal, we developed a 3D brain registration method, called HAMMER, for inter-subject registration, and also a 4D (3 spatial dimensions + 1 temporal dimension) brain registration method to obtain more accurate measurements for tiny longitudinal brain changes, compared to the 3D registration methods. For better alignment of a population of images, we have recently developed several groupwise registration methods for joint registration of all images together, thus further improving the registration accuracy among all images. With accurate brain structural/functional information measured by our registration methods, we further developed a multivariate analysis method, based on support vector machine, to jointly consider all structural/functional changes for determining the group difference between normal and abnormal brains, i.e., due to diseases, aging, or brain disorders. We have applied our developed methods to computer-aided diagnosis of schizophrenia and AD. Details of these 3D, 4D, and groupwise registration methods, as well as brain classification methods, will be discussed in this talk. |
