Identification or counting of sprites in a picture usually requires the use of shape descriptors. The picture has already been segmented, through any common method – the most simple being a mere segmentation by threshold. This site is about Fourier descriptors, which give good results with the complex method for describing closed curves.
Such a technique is commonly used for pattern recognition like chromosome classification, identification of aircrafts or identification of particules.
A big issue about Fourier descriptors is how many terms should be kept
from the Fourier transform so the description is efficient. After introducing
the two methods and pointing out the advantages of the complex method,
we will describe an algorithm that allows to test it according to the number
of Fourier descriptors, and then give our results and conclusions. We will
eventually give two examples of less common applications for Fourier descriptors.