In a neuro-radiological study, we obtain an image of the human brain. In order to determine if there is a pathology or not, we usually use a comparison with a normal sample. and this sample can be provided by an atlas of normal anatomy. In order to compare the sample and the patient's brain, it is required to introduce some coordinate system in order to co-register objects to each other, otherwise the comparison would not have any meaning.
A Simple and Fast Method of 3D Registration and Statistical Landmark Localization for Sparse Multi-Modal/Time-Series Neuroimages Based on Cortex Ellipse Fitting
Usually there are several ways to perform this mapping: either to use cross-correlation in order to find the best nonlinear correspondence between the Atlas and the brain, or identify and locate some spacial structures and at next to choose a set of points that will participate in the mapping of the both objects.
We propose to introduce a special coordinate system for the brain that is based on the shape of the brain. We showed that if we considered the shape of the brain around the midsagittal plane, then some of the brain structure, such as AC and PC can be located statistically with a fairly good accuracy. It also allows us to map the brains in time series to find a corresponding brain structures at different periods of time, as well as to see the dynamics of changes. This can be used to find a match between images of the brain taken at different time intervals and using different modalities. This introduction of coordinate system allows us to introduce an intrinsic coordinate system associated with the shape of the brain.