Object recognition using an affine invariant wavelet representation

A novel algorithm based on the dyadic wavelet transform for recognising a planar object undergoing a general affine transformation is presented. The proposed algorithm has two steps: constructing the representation and the matching process. Two different wavelets associated with a given scaling function are used in the construction. In the matching procedure, only extrema of the representation are used. This makes the process efficient and less sensitive to small variations in the representation. Experimental results show that the representation is robust and, combined with the matching algorithm, it efficiently classifies unknown objects