ABSTRACT
In many applications, such as medical imaging, data are acquired in a discrete form and approximations are needed to obtain a continuous model. There fore, it is suitable for the algorithm, called here curve algorithm, to be in dependent of the points location. Euclidean invariance is widely recognized as an important property of curve and surface shape description algorithms for computer-aided design, pattern recognition and image analysis. The Eu clidean transformations are defined by combining rotations and translations, and the Euclidean invariance is equivalent to axis-independence.
