An optimal feature extraction method

The authors present a feature extraction method that can find a minimum set of minimal features for any given set of N binary images (patterns). This method has some similarities to the method of minimizing multiple switching functions. It becomes difficult to apply manually as N becomes large. For large N, a computer implementation of this method is presented. Two examples, one for N=5 (small N) and the other for N=10 (large N), are given to illustrate the method. This method can easily be extended and applied to optimal feature extraction of high-dimensional multigrey color images.<
>