Fusion of Iris and Periocular Biometrics for Cross-Sensor Identification

As a reliable personal identification method, iris recognition has been widely used for a large number of applications. Since a variety of iris devices produced by different vendors may be used for some large-scale applications, it is necessary to match heterogeneous iris images against the variations of sensors, illuminators, imaging distance and imaging conditions. This paper aims to improve cross-sensor iris recognition performance using a novel multi-biometrics strategy. The novelty of our solution is that both iris and periocular biometrics in heterogeneous iris images are combined through score-level information fusion for approaching the problem of iris sensor interoperability. Then the improved feature extraction method, namely Multi-Directions Ordinal Measures, is applied to encode both iris and periocular images to describe the distinctive features. The experimental results on images captured from three iris devices, including two close-range iris devices and one long-range iris device, demonstrate the effectiveness of the proposed method.