A measure of information gained through biometric systems

Takahashi, Kazutoshi; Murakami, Takao

doi:10.1016/j.imavis.2013.12.010

Cited by 30 publications

(49 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, it is not straightforward to obtain a precise estimate of individuality of the IrisCode representation using the above result because it fails to take into account the genuine score distribution (consequently, intra-subject variations are not modeled). A simple extension of the above approach is to measure the relative entropy between genuine and impostor match score distributions [45]. But this approach may grossly underestimate the entropy of the biometric features and the resulting entropy estimates should be considered as a very loose lower bound.…”

Section: ) Biometric Entropy Estimationmentioning

confidence: 99%

Biometric Template Protection: Bridging the performance gap between theory and practice

Nandakumar

Jain²

2015

IEEE Signal Process. Mag.

265

147

View full text Add to dashboard Cite

Abstract-Biometric recognition is an integral component of modern identity management and access control systems. Due to the strong and permanent link between individuals and their biometric traits, exposure of enrolled users' biometric information to adversaries can seriously compromise biometric system security and user privacy. Numerous techniques have been proposed for biometric template protection over the last 20 years. While these techniques are theoretically sound, they seldom guarantee the desired non-invertibility, revocability, and non-linkability properties without significantly degrading the recognition performance. The objective of this work is to analyze the factors contributing to this performance gap and highlight promising research directions to bridge this gap. Design of invariant biometric representations remains a fundamental problem, despite recent attempts to address this issue through feature adaptation schemes. The difficulty in estimating the statistical distribution of biometric features not only hinders the development of better template protection algorithms, but also diminishes the ability to quantify the non-invertibility and non-linkability of existing algorithms. Finally, achieving nonlinkability without the use of external secrets (e.g., passwords) continues to be a challenging proposition. Further research on the above issues is required to cross the chasm between theory and practice in biometric template protection.

show abstract

Section: ) Biometric Entropy Estimationmentioning

confidence: 99%

Biometric Template Protection: Bridging the performance gap between theory and practice

Nandakumar

Jain²

2015

IEEE Signal Process. Mag.

265

147

View full text Add to dashboard Cite

show abstract

“…In [20], [21], a biometric information measure based on relative entropy is proposed. However, in this case, biometric information measure considers inter-user distance and intrauser distance distributions rather than using feature distributions directly.…”

Section: B Relative Entropymentioning

confidence: 99%

“…For example, while the biometric information measured for face images using effective keyspace (as introduced in [15]) is about 2.6 bits, biometric feature information of face for different feature representations is found to be 35 ∼ 55 bits [4]. On the other hand, in [21], authors reported the biometric system entropy for face is about 12.6 bits. It can be seen that biometric information measured using various methods reported in literature are not only different, but also their findings vary widely.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

What is biometric information and how to measure it?

Sutcu

Tabassi

Sencar

et al. 2013

2013 IEEE International Conference on Technologies for Homeland Security (HST)

View full text Add to dashboard Cite

Being able to measure the information content of biometric data has significant importance. However, it is not only difficult to define the biometric information conceptually, but also a challenging estimation problem due to many practical issue such as, different feature representation of different biometric modalities, unknown feature distributions, limited sample size, etc.In this paper, we provided a summary of the several proposed approaches and discuss their merits and shortcomings of these works and focused on techniques which are based on relative entropy. We also introduced a user-specific biometric information measure and evaluated these measures using iris biometric. We explored the effect of the biometric sample quality on the biometric information measurements using different subsets of the iris data with different quality levels. We also investigated the effect of binarization/encoding phase in the iris recognition on the biometric information by comparing our results obtained from image and binary feature domain.

show abstract

“…Kenta Takahashi, et al, [6] presented an article on a measure of information gained through biometric matching systems. In this paper, they discussed how the information about the identity of an individual is obtained from biometric samples through the biometric systems and how to evaluate the Biometric System Entropy based on the mutual information.…”

Section: Iintroductionmentioning

confidence: 99%

Efficient personal identification using multimodal biometrics

Ahilandeswari¹,

Prabu²,

Priyadharshini³

et al. 2015

2015 International Conference on Circuits, Power and Computing Technologies [ICCPCT-2015]

View full text Add to dashboard Cite

For recent years, the use of personal identity systems using multimodal biometrics has been increasing tremendously in number of fields. Although the Unimodal biometric systems serve well in various areas, it is notable that they have disadvantages regarding security and accuracy. Multimodal biometric systems focus on combining more than one possible biometric technology in order to secure the applications to a great extent. This in turn resulted in the improvement of performance and robustness against fraudulent attacks. Personal identification plays a major role in any information sharing process. For this, the identification systems must be designed in such a manner that it should minimize the system error rates, susceptibility mimics and false match rate. Considering these factors, we aim at providing a new personal identification system using multimodal biometrics. The proposed system uses three of the biometric technologies in the process of identification: (i) Face recognition; (ii) Fingerprint recognition; and (iii) Speech Recognition. Here we discuss about the methods of feature extraction, fusion and decision used in our system and their advantages over other biometric systems.

show abstract

A measure of information gained through biometric systems

Cited by 30 publications

References 7 publications

Biometric Template Protection: Bridging the performance gap between theory and practice

Biometric Template Protection: Bridging the performance gap between theory and practice

What is biometric information and how to measure it?

Efficient personal identification using multimodal biometrics

Contact Info

Product

Resources

About