|
The problem of matching 2D and 3D elements is an issue that affects many fields of science, such
as archaeology or medicine. The article concerns devising an original method of constructing
effective algorithms for matching 2D elements that would allow to automate the process of arranging
broken archaeological and museum elements.
The described method of matching 2D elements is based on the contours data in the form of string
codes, where each character reflects both length and direction. Linguistic metrics were used to
compare the elements. Furthermore, the developed algorithms perform an exhaustive recursive
search. The results of numerical experiments showed that the time of calculations, and thus the
number of potential element matching, increases along with the number of elements and with the
length of the compared substrings. Additionally, it depends on the uniqueness of the fragments of
the contours of each element.
The applied optimizations which included changing the length of the compared substrings
and replacing elements random sampling with ordered sampling, resulted in both a reduction
in computation time as well as increasing the knowledge about the functioning of the method
and algorithm.
|
