New approach for 3D object recognition using non uniform fourier and moments coefficients

In this paper,new descriptors have been proposed which are invariant to rotation and affinity transformations (non uniform fourier and moments coefficients). these descriptors are computed using fourier or moments coefficients obtained from development of object coordinates. these coordinates are parameterized with a function also invariant to affinity. the normal fourier transform depends on points index,the non uniform transform proposed here depend on a parameter. this parameter depends on shape not on order of points. the proposed descriptors are easy to extract and are extensible to higher dimension. © school of engineering,taylor’s university.