A comparative study of moment invariants and Fourier descriptors in planar shape recognition

Recent developments in the area of pattern recognition have brought up various methodologies to the problem of planar shape recognition. In this study, two popular feature sets, namely moment invariants and Fourier descriptors are applied to the problem of classification of 2-D images of airplanes, and their performance is compared vis-a-vis computational load and accuracy. An alternative form of the moment invariant technique, Zernike moments, is also compared with ordinary moment invariants.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>