Achieving privacy-preserving iris identification via El Gamal

Currently, many biometric systems maintain the user’s biometrics and templates in plaintext format, which brings great privacy risk to uses’ biometric information. Biometrics are unique and almost unchangeable, which means it is a great concern for users on whether their biometric information would be leaked. To address this issue, this paper proposes a confidential comparison algorithm for iris feature vectors with masks, and develops a privacy-preserving iris verification scheme based on the El Gamal encryption scheme. In our scheme, the multiplicative homomorphism of encrypted features is used to compare of iris features and their mask information. Also, this paper improves the Hamming distance of iris features, which makes the similarity matching work better than existing ones. Experimental results confirm the practicality of our proposed schemes in real world applications, that is, for the iris feature vectors and masks of 2048 bits, nearly 12 comparisons can be performed per second.