PDF version of this document
After applying some changes to the code and running some over-night experiments, the code could finally be used in a way which made it usable and effective. At first, the original bump data was used with varying levels of variation that is inherent in the data. At a later stage, hand data was used as well to test a more realistic and interesting example.
Once the code was in full working order and an interface was in place, experiments could be set up and run very rapidly. The rest of this section describes some of these experiments and their corresponding and related equivalents. Many of the same ideas were parallelly applied to the code handling images and their model-based registration, namely that of AART.