Gender recognition using fisherfaces and a fuzzy iterative self-organizing technique

This paper proposes a new gender recognition method by employing Fisherfaces and the fuzzy iterative self-organizing technique (ISODATA). The proposed method first uses Fisherfaces to extract suitable features from the reduced dimensional space. Then, the optimal fuzzy cluster centers can be calculated by applying the fuzzy ISODATA model to learn and cluster the gender features. Finally, the fuzzy nearest-neighbor is used for classification. The proposed method inherits the advantages of Fisherfaces and the fuzzy ISODATA method, which can extract suitable features for recognition and obtain the best clustering centers without the need for priori. Experimental results show the proposed method outperforms the mainstream methods in recognition rate and testing time.