Cross-equipment biometric feature recognition method

The invention discloses a cross-equipment biometric feature recognition method. The method comprises the following steps: biometric feature image groups acquired by different equipment are preprocessed, and local subarea features of every biometric feature image are extracted; related features are selected with the similarity as guidance; the related features are trained to obtain classifier models; the preprocessing and the related feature extraction are implemented for the biometric feature images to be tested from the different acquiring equipment; the related features are respectively input in the classifier models for classifying; the biometric feature images from the different acquiring equipment are compared by the classifying results of the classifier models to finally obtain the cross-equipment biometric feature recognition results. The selected features can preferably overcome the noise brought by the cross-equipment image comparison to improve the cross-equipment biometric feature image recognition performance; the cross-equipment biometric feature recognition method can be applied to large-scale identity authentication systems for biometric feature recognition and other application systems needing security defense.