Novel Methods for the Visualization and Analysis of Functional Maps in Cortex

Methods 92 the only way to restrict the area where the values, which end up at the control points, are in- terpolated. Figure 44 shows a definition that is less restrictive. There are many more possibili- ties. Since the definition in Figure 43 yields good results, was easy to implement, and fast to calculate, it is used in the current implementation of mrFindBorders. 5.2.8 Hierarchical Matching The matching algorithm is applied iteratively in a multi-level or multi-scale hierarchy. This paradigm is often used, for example in multigrid methods. At first, the matching is done on low-resolution images with only a few control points. The early fit is therefore obtained quickly. Then, increasing the image resolution and the number of control points so that finer (9) (9’) (7’) (6’) (5’) (7) (6) (5) x y (1’) (2’) (3’) (4’) (8’) (B) Figure 44: Constraints for displacement values in 2D See text for details