Abstract-In this paper, we propose a ranking based locality sensitive hashing inspired two-factor cancelable biometrics, dubbed "Index-of-Max" (IoM) hashing for biometric template protection. With externally generated random parameters, IoM hashing transforms a real-valued biometric feature vector into discrete index (max ranked) hashed code. We demonstrate two realizations from IoM hashing notion, namely Gaussian Random Projection based and Uniformly Random Permutation based hashing schemes. The discrete indices representation nature of IoM hashed codes enjoy several merits. Firstly, IoM hashing empowers strong concealment to the biometric information. This contributes to the solid ground of non-invertibility guarantee. Secondly, IoM hashing is insensitive to the features magnitude, hence is more robust against biometric features variation. Thirdly, the magnitude-independence trait of IoM hashing makes the hash codes being scale-invariant, which is critical for matching and feature alignment. The experimental results demonstrate favorable accuracy performance on benchmark FVC2002 and FVC2004 fingerprint databases. The analyses justify its resilience to the existing and newly introduced security and privacy attacks as well as satisfy the revocability and unlinkability criteria of cancelable biometrics.Index Terms-Fingerprint, cancelable template, Index-of-Max hashing, security and privacy I. INTRODUCTIONATELY, rapid proliferation of biometric applications leads to massive amount of biometric templates. Public worries about the security and privacy of the biometric templates if stolen or compromised. Such concerns are attributed to the strong binding of individuals and privacy, and further complicated by the fact that biometric traits are irrevocable. Given the above threats, a number of proposals have been reported to protect the biometric templates. Generally, the proposals available in the literature can be broadly divided into two categories: feature transformation (or cancelable biometrics) and biometric cryptosystems (biometric encryption). Biometric cryptosystem serves the purpose of either securing the secret using biometric feature (key binding) or generating the secret directly from the biometric feature (key generation). On the other hand, cancelable biometrics [1] is truly meant for biometric template protection. It refers to the irreversible transform that can alter the biometric templates such that security and privacy of the templates can be assured. If a cancelable template is compromised, a new template can be re-generated from the same biometrics.The cancelable biometric schemes in the literature vary according to different biometric modalities. However, the operation of a general cancelable biometrics system is similar to the conventional biometric system where the system composes of sensor, feature extractor and matcher except the former includes a parameterized transformation function right after feature extractor, and the matching is done in the transformed domain, rather in the feature doma...

No abstract

No abstract

Secure sketch conceals any random string w by generating a helper string ss (known as a sketch). It allows the exact recovery of w from ss given another value w that is close to w. A secure sketch can be utilized to protect any error-prone secret, e.g., biometrics, stored in secret storage to promote secure authentication. When error tolerance is demanded, a secure sketch can be used as an error correction code to tolerate the noise over an unreliable, noisy communication channel. However, when both security and error tolerance are of interest, the error tolerance property of a secure sketch imposes entropy loss. It leads to a weak security guaranty on a low entropy input string. Recent work by Fuller et al. (2016) has exploited the structure of the input string. They showed that having precise knowledge over the input strings' distribution is essential to construct a secure sketch for an input string of low entropy. We formalized a new model for secure sketch construction to realize precise knowledge of the input distribution setting. With the formalized new model, we devised an explicit secure sketch construction to a large family of noisy sources. The devised secure sketch can tolerate an error rate close to 1/2 in polynomial time, i.e., O(n 4 ), and meets the best possible secure sketch's security bound with optimal entropy loss.

No abstract