Abstract
In this paper we present a novel hand posture recognizer based on wavelet network learnt by fast wavelet transform (FWN) including a fuzzy decision support system (FDSS). Our contribution in this paper resides in proposing a new classification way for the FWN classifier. The FWN having an hybrid architecture (using as activation functions both wavelet and scaling ones) provides hybrid weight vectors when approximating an image. The FWN classification phase was achieved by computing simple distances between test and training weight vectors. Those latter are composed of two types of coefficients that are not in the same value range which may influence on the distances computing. This can cause wrong recognitions. So, to overcome this lacuna, a new classification strategy is proposed. It operates a human reasoning mode employing a FDSS to calculate similarity degrees between test and training images. Comparisons with other works are presented and discussed. Obtained results have shown that the new hand posture recognizer performs better than previously established ones. ...