A Multimodal Gender Recognition Based on Bayesian Hierarchical Model

In the field of gender recognition, face vision information is an important factor. But to achieve strong robustness and high recognition performance, a new framework fusing the fingerprint and face image representation is put forward in this paper. At the stage of image feature representation, "bag of words model" is used to capture the significant features in fingerprint and face image. In the decision making layer, the recognition results can be obtained by fusing gender estimates in different modals. A Multimodal gender recognition Based on Bayesian hierarchical model is experimented on the fingerprint and face image database. The effectiveness of the new framework fusing fingerprint and face image information is verified and the feature representation and generative modal in this paper are both effective.