Ethnicity classification based on fusion of face and gait

The recognition of ethnicity of an individual can be very useful in a video-based surveillance system. In this paper, we propose a multimodal biometric system involving an integration of frontal face and lateral gait, for the specific problem of ethnicity classification. This system performs a feature fusion to improve the discrimination of human ethnicity. Face features are extracted by means of the uniform LBP operator and gait information is characterized by a spatio-temporal representation. Afterwards, canonical correlation analysis (CCA), as a powerful tool to relate two sets of measurements, is used to fuse the two modalities at the feature level. A database including 36 walking people from East Asia and South America is built for the purpose of ethnicity classification. The experimental results show that the ethnicity recognition rate is improved by fusing face and gait information.