FMR image registration using the Mumford-Shah functional and Shape information

Motion can have a significant impact on signal changes in functional MR images, and affect the detection of real functional activation. To minimize the effect of motion on the fMRI signal, we propose a simultaneous segmentation and registration model. Due to T2 weighted signal loss, decreased resolution, and low contrast in fMR images, the feature to be used for realignment can't be detected reliably by using only image information. Our approach uses prior shape information in the Mumford-Shah segmentation [MS] technique to find the feature in a lowresolution image. In addition to this, the spatial transform that maps the segmented contour to the given shape is found. Given a time serier of images, we use this method to find transformations to realign the images. In this model, we minimize an energy functional that depends on the Mumford-Shah energy functional and the shape of interest, so that the boundary of the object can be captured by the different intensity distributions inside and outside the feature and the prior knowledge of its shape. The model has been tested both on synthetic data and fMR brain image data. The experimental results showed the effectiveness of this model in feature determination and in time series image registration.