Multimodal Template Protection Based on Data Transformation and Fuzzy Vault ⋆

Due to the similarity of biological characteristics in image acquisition and physiological complementarity between the face and the ear, this paper proposes a hybrid face and ear multi-biometric template protection method, which firstly conducts a feature level fusion of the face and the ear feature vectors, and secondly transforms the fused features with a non-invertible function, and then utilize Fuzzy Vault algorithm to protect the transformed templates. The basic flow of the scheme is divided into several parts: image preprocessing, Gabor-PCA feature extraction, feature level fusion, template transformation, fusion template encryption, fusion template decryption. On two face and ear multimodal datasets, the face and ear authentication experimental results show the effectiveness of the proposed multimodal template protection method.