$next$ $up$ $previous$ $contents$ $index$
Next: New Algorithms Up: ALGORITHMS Previous: Main Loop Contents Index

Algorithm Visualised

Figure $[*]$ shows what is done in the algorithm above in a very loose and simplified form. A reference image, as shown at the top stays unaffected, while all other images are manipulated in the way described in Figure $[*]$ . These are used to construct a model which then infers a certain complexity measures, e.g. description length (as in Chapter $[*]$ ). Based on that measures of complexity, subsequent warps are applied to the group of images.

**Figure:** Schematic of the current registration algorithm. A reference image and the rest of the warped set form a combined model which is evaluated in an MDL-like manner.

**Figure:** Current algorithm at a lower level. The idea of a reparameterisation is shown by emphasising that images are formed by aggregation of the previous image with some parameterisation.

For a full corresponding notation and further explanation on the figures, see the short presentation file at:

http://www.danielsorogon.com/Webmaster/Research/Explanatory_Notes/2004/Group-wise_NRR_Strategy_and_Notation.pdf

It is certainly beyond the scope of this report and can be remitted. However, some of the figures should be thought-inspiring. A brief explanation of each has been included above for this reason.

$next$ $up$ $previous$ $contents$ $index$
Next: New Algorithms Up: ALGORITHMS Previous: Main Loop Contents Index

2004-08-02