QFuse: Online learning framework for adaptive biometric system

doi:10.1016/J.PATCOG.2015.05.002

Home
/
Papers
/
QFuse: Online learning framework for adaptive biometric system

Journal Article•DOI•

QFuse: Online learning framework for adaptive biometric system

Samarth Bharadwaj¹, Himanshu Bhatt¹, Richa Singh¹, Mayank Vatsa¹, Afzel Noore¹ - Show less +1 more•Institutions (1)

Indraprastha Institute of Information Technology¹

01 Nov 2015-Pattern Recognition (Pergamon)-Vol. 48, Iss: 11, pp 3428-3439

TL;DR: This paper presents an adaptive context switching algorithm coupled with online learning to address the scalability and accommodate the variations in data distribution of biometrics.

read less

About: This article is published in Pattern Recognition.The article was published on 2015-11-01. It has received 27 citations till now. The article focuses on the topics: Biometrics.

...read moreread less

Citations

PDF

Open Access

More filters

Journal Article•DOI•

A comprehensive overview of biometric fusion

[...]

Maneet Singh¹, Richa Singh¹, Arun Ross²•Institutions (2)

Indraprastha Institute of Information Technology¹, Michigan State University²

01 Dec 2019-Information Fusion

TL;DR: A comprehensive review of techniques incorporating ancillary information in the biometric recognition pipeline is presented in this paper, where the authors provide a comprehensive overview of the role of information fusion in biometrics.

...read moreread less

151 citations

Journal Article•DOI•

Overview of the combination of biometric matchers

[...]

Alessandra Lumini¹, Loris Nanni²•Institutions (2)

University of Bologna¹, University of Padua²

01 Jan 2017-Information Fusion

TL;DR: Several systems and architectures related to the combination of biometric systems, both unimodal and multimodal, are overviews, classifying them according to a given taxonomy, and a case study for the experimental evaluation of methods for biometric fusion at score level is presented.

...read moreread less

123 citations

Cites background or methods from "QFuse: Online learning framework fo..."

...Reference Name Year Biometric traits Users [61] WVU 2007 Fingerprint, Face, Iris, Palmprint, Hand geometry, Voice 270 [62] MBGC 2009 Face, Iris >146 [63] BiosecurID 2010 Face, Speech, Iris, Signature, Handwritting, Fingerprints, Hand, Keystroking 400 [60] BMDB 2010 Face, Speech, Signature, Fingerprints, Hand, Iris >600 [64] SDUMLA-HMT 2011 Face, Gait, Iris, Fingerprints 106 [65] MOBio 2012 Face, Speech 152 [66] MMU GASPFA 2013 Gait, Speech, Face 82 [67] MobBIO 2014 Face, Iris, Voice 105 [59] gb2sμMOD 2015 Hand, Iris and Face 60 [40] LEA 2015 Fingerprint, Face, Iris 18,000 Table 2....
[...]
...[60] BMDB 2010 Face, Speech, Signature, Fingerprints, Hand, Iris >600 [64] SDUMLA-HMT 2011 Face, Gait, Iris, Fingerprints 106 [65] MOBio 2012 Face, Speech 152 [66] MMU GASPFA 2013 Gait, Speech, Face 82 [67] MobBIO 2014 Face, Iris, Voice 105 [59] gb2sμMOD 2015 Hand, Iris and Face 60 [40] LEA 2015 Fingerprint, Face, Iris 18,000 Table 2....
[...]
...LEA [40] is a multimodal database provided by a Law Enforcement Agency and captured in unconstrained real world conditions with uncooperative users....
[...]
...QFuse [40] is an online learning algorithm for adaptive biometric fusion that incorporates image quality in the dynamic selection of unimodal classifiers and their fusion....
[...]
...To deal with this problem, combination methods based on online learning [40] are required, which provide an efficient way to sustain the performance by addressing the variations in data (match score and quality score) distribution introduced by the newly enrolled individuals....
[...]

Journal Article•DOI•

EmbraceNet: A robust deep learning architecture for multimodal classification

[...]

Jun-Ho Choi¹, Jong-Seok Lee¹•Institutions (1)

Yonsei University¹

01 Nov 2019-Information Fusion

TL;DR: In this article, a deep learning-based multimodal fusion architecture for classification tasks is proposed, which guarantees compatibility with any kind of learning models, deals with cross-modal information carefully, and prevents performance degradation due to partial absence of data.

...read moreread less

82 citations

Journal Article•DOI•

Group sparse representation based classification for multi-feature multimodal biometrics

[...]

Gaurav Goswami¹, Paritosh Mittal¹, Angshul Majumdar¹, Mayank Vatsa¹, Richa Singh¹ - Show less +1 more•Institutions (1)

Indraprastha Institute of Information Technology¹

01 Nov 2016-Information Fusion

TL;DR: The Group Sparse Representation based Classifier (GSRC) is proposed which removes the requirement for a separate feature-level fusion mechanism and integrates multi-feature representation seamlessly into classification.

...read moreread less

63 citations

Cites background or methods from "QFuse: Online learning framework fo..."

...8% is obtained using the context switching algorithm [37]....
[...]
...16 (a) 1 2 3 4 5 20 25 30 35 40 45 50 Rank Id e n ti fi c a ti o n A c c u ra c y ( % ) Face UCLBP Face SURF SRC Face GSRC Face 1 2 3 4 5 40 45 50 55 60 65 70 Rank Id e n ti fi c a ti o n A c c u ra c y ( % ) Finger NBIS Finger VF SRC Finger GSRC Finger (b) 1 2 3 4 5 30 35 40 45 50 55 Rank Id e n ti fi c a ti o n A c c u ra c y ( % ) Iris VASIR Iris LPG SRC Iris GSRC Iris (c) Figure 5: CMC curves for identification on the LEA multimodal database: individual features, SRC and GSRC on (a) face, (b) fingerprint,and (c) iris....
[...]
...The 18 (a) (b) 1 2 3 4 5 93 94 95 96 97 98 99 100 Rank Id e n ti fi c a ti o n A c c u ra c y ( % ) SRC Face Iris SRC Iris Finger GSRC Face Iris GSRC Face Finger GSRC Iris Finger 1 2 3 4 5 95 95.5 96 96.5 97 97.5 98 98.5 99 99.5 100 Rank Id e n ti fi c a ti o n A c c u ra c y ( % ) Sum Rule Fusion SRC Combined Context Switching GSRC Combined SRC Face Finger Figure 6: CMC curves for identification on the WVU multimodal database....
[...]
...Identification experiments are performed on both the WVU and LEA databases and the performance of the proposed GSRC algorithm is evaluated in four scenarios and major observations are noted below....
[...]
...On the LEA database, the context switching algorithm outperforms both sum rule fusion and traditional SRC with an identification performance of 55.8% whereas the GSRC algorithm performs 6.5% better and achieves 62.3% rank-1 accuracy....
[...]

Journal Article•DOI•

Liveness Detection and Automatic Template Updating Using Fusion of ECG and Fingerprint

[...]

Majid Komeili¹, Narges Armanfard², Dimitrios Hatzinakos¹•Institutions (2)

University of Toronto¹, McMaster University²

12 Feb 2018-IEEE Transactions on Information Forensics and Security

TL;DR: This paper combines ECG with a fingerprint liveness detection algorithm and proposes a stopping criterion that reduces the average waiting time for signal acquisition and examines automatic template updating using ECG and fingerprint.

...read moreread less

Abstract: Fingerprints have been extensively used for biometric recognition around the world. However, fingerprints are not secrets, and an adversary can synthesis a fake finger to spoof the biometric system. The mainstream of the current fingerprint spoof detection methods are basically binary classifier trained on some real and fake samples. While they perform well on detecting fake samples created by using the same methods used for training, their performance degrades when encountering fake samples created by a novel spoofing method. In this paper, we approach the problem from a different perspective by incorporating electrocardiogram (ECG). Compared with the conventional biometrics, stealing someone’s ECG is far more difficult if not impossible. Considering that ECG is a vital signal and motivated by its inherent liveness, we propose to combine it with a fingerprint liveness detection algorithm. The combination is natural as both ECG and fingerprints can be captured from fingertips. In the proposed framework, the ECG and fingerprint are combined not only for authentication purpose but also for liveness detection. We also examine automatic template updating using ECG and fingerprint. In addition, we propose a stopping criterion that reduces the average waiting time for signal acquisition. We have performed extensive experiments on the LivDet2015 database which is presently the latest available liveness detection database and compare the proposed method with six liveness detection methods as well as 12 participants of LivDet2015 competition. The proposed system has achieved a liveness detection equal error rate (EER) of 4.2% incorporating only 5 s of ECG. By extending the recording time to 30 s, liveness detection EER reduces to 2.6% which is about 4 times better than the best of six comparison methods. This is also about 2 times better than the best results achieved by the participants of the LivDet2015 competition.

...read moreread less

48 citations

Cites background from "QFuse: Online learning framework fo..."

...While multimodal biometric systems based on conventional traits such as face and fingerprint have been extensively investigated in [32] and [33], there exist only a few works about a multimodal biometric system that includes ECG....
[...]

1
2
3
4
…
5
6

Collapse

References

PDF

Open Access

More filters

Book•DOI•

Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond

[...]

Bernhard Schölkopf¹, Alexander J. Smola•Institutions (1)

Max Planck Society¹

01 Dec 2001

TL;DR: Learning with Kernels provides an introduction to SVMs and related kernel methods that provide all of the concepts necessary to enable a reader equipped with some basic mathematical knowledge to enter the world of machine learning using theoretically well-founded yet easy-to-use kernel algorithms.

...read moreread less

Abstract: From the Publisher: In the 1990s, a new type of learning algorithm was developed, based on results from statistical learning theory: the Support Vector Machine (SVM). This gave rise to a new class of theoretically elegant learning machines that use a central concept of SVMs-kernels--for a number of learning tasks. Kernel machines provide a modular framework that can be adapted to different tasks and domains by the choice of the kernel function and the base algorithm. They are replacing neural networks in a variety of fields, including engineering, information retrieval, and bioinformatics. Learning with Kernels provides an introduction to SVMs and related kernel methods. Although the book begins with the basics, it also includes the latest research. It provides all of the concepts necessary to enable a reader equipped with some basic mathematical knowledge to enter the world of machine learning using theoretically well-founded yet easy-to-use kernel algorithms and to understand and apply the powerful algorithms that have been developed over the last few years.

...read moreread less

7,880 citations

Journal Article•DOI•

On combining classifiers

[...]

Josef Kittler¹, M. Hatef², Robert P. W. Duin³, Jiri Matas¹•Institutions (3)

University of Surrey¹, ERA Technology Ltd², Delft University of Technology³

01 Mar 1998-IEEE Transactions on Pattern Analysis and Machine Intelligence

TL;DR: A common theoretical framework for combining classifiers which use distinct pattern representations is developed and it is shown that many existing schemes can be considered as special cases of compound classification where all the pattern representations are used jointly to make a decision.

...read moreread less

Abstract: We develop a common theoretical framework for combining classifiers which use distinct pattern representations and show that many existing schemes can be considered as special cases of compound classification where all the pattern representations are used jointly to make a decision. An experimental comparison of various classifier combination schemes demonstrates that the combination rule developed under the most restrictive assumptions-the sum rule-outperforms other classifier combinations schemes. A sensitivity analysis of the various schemes to estimation errors is carried out to show that this finding can be justified theoretically.

...read moreread less

5,670 citations

"QFuse: Online learning framework fo..." refers methods in this paper

...In matcher fusion, all the constituent matchers are used and their evidences are combined using fusion rules [4], [5], [6], [7]....
[...]

Journal Article•DOI•

Face Description with Local Binary Patterns: Application to Face Recognition

[...]

Timo Ahonen¹, Abdenour Hadid¹, Matti Pietikäinen¹•Institutions (1)

University of Oulu¹

01 Dec 2006-IEEE Transactions on Pattern Analysis and Machine Intelligence

TL;DR: This paper presents a novel and efficient facial image representation based on local binary pattern (LBP) texture features that is assessed in the face recognition problem under different challenges.

...read moreread less

Abstract: This paper presents a novel and efficient facial image representation based on local binary pattern (LBP) texture features. The face image is divided into several regions from which the LBP feature distributions are extracted and concatenated into an enhanced feature vector to be used as a face descriptor. The performance of the proposed method is assessed in the face recognition problem under different challenges. Other applications and several extensions are also discussed

...read moreread less

5,563 citations

"QFuse: Online learning framework fo..." refers methods in this paper

...To match two corresponding UCLBP features or SURF descriptors, χ2 distance measure is used....
[...]
...For face, two matchers are used: Uniform Circular Local Binary Pattern (UCLBP) [44] as face matcher1 and Speeded Up Robust Features (SURF) [45] as face matcher2....
[...]
...UCLBP is computed with circular encoding of eight neighboring pixels evenly positioned on a circle of radius two....
[...]

Journal Article•DOI•

How iris recognition works

[...]

John Daugman¹•Institutions (1)

University of Cambridge¹

01 Jan 2004-IEEE Transactions on Circuits and Systems for Video Technology

TL;DR: Algorithms developed by the author for recognizing persons by their iris patterns have now been tested in many field and laboratory trials, producing no false matches in several million comparison tests.

...read moreread less

Abstract: Algorithms developed by the author for recognizing persons by their iris patterns have now been tested in many field and laboratory trials, producing no false matches in several million comparison tests. The recognition principle is the failure of a test of statistical independence on iris phase structure encoded by multi-scale quadrature wavelets. The combinatorial complexity of this phase information across different persons spans about 249 degrees of freedom and generates a discrimination entropy of about 3.2 b/mm/sup 2/ over the iris, enabling real-time decisions about personal identity with extremely high confidence. The high confidence levels are important because they allow very large databases to be searched exhaustively (one-to-many "identification mode") without making false matches, despite so many chances. Biometrics that lack this property can only survive one-to-one ("verification") or few comparisons. The paper explains the iris recognition algorithms and presents results of 9.1 million comparisons among eye images from trials in Britain, the USA, Japan, and Korea.

...read moreread less

2,829 citations

Proceedings Article•DOI•

How iris recognition works

[...]

John Daugman¹•Institutions (1)

University of Cambridge¹

10 Dec 2002

...read moreread less

Abstract: The principle that underlies the recognition of persons by their iris patterns is the failure of a test of statistical independence on texture phase structure as encoded by multiscale quadrature wavelets. The combinatorial complexity of this phase information across different persons spans about 249 degrees of freedom and generates a discrimination entropy of about 3.2 bits/mm/sup 2/ over the iris, enabling real-time decisions about personal identity with extremely high confidence. Algorithms first described by the author in 1993 have now been tested in several independent field trials and are becoming widely licensed. This presentation reviews how the algorithms work and presents the results of 9.1 million comparisons among different eye images acquired in trials in Britain, the USA, Korea, and Japan.

...read moreread less

2,437 citations

"QFuse: Online learning framework fo..." refers methods in this paper

...A 8×8 convolution filter from the pre-processing step of Daugman [47] is used to measure the 2D spectral power of the image as given by Equation A....
[...]
...[47] J. Daugman, How iris recognition works, IEEE Transactions on Circuits and Systems for Video Technology 14 (11) (2004) 21–30....
[...]
...A 8×8 convolution filter from the pre-processing step of Daugman [47] is used to measure the 2D spectral power of the image as given by Equation A.9. f(x) = 100 x2 (x2 + c2) (A.9) where x is the total spectral power measured by the convolution filter and c is the half power, corresponding to 50%....
[...]