Age Bias in Finger Vein Biometric Research

Joanne L. Hall

1 a

, Jomin John

, Jessica Liebig

2 b

and Anju Skariah

School of Science, RMIT University, Melbourne, Australia

Health and Biosecurity, Commonwealth Scientiﬁc and Industrial Research Organisation, Brisbane, Australia

Keywords:

Finger Vein Biometrics, Vein Patterns, Datasets, Age Bias.

Abstract:

Finger vein biometrics have been implemented for authentication in a variety of contexts and places. Vein

patterns are unique, easy to capture and resistant to surface wear and tear. However, there has been a lack

of research on the effectiveness and stability of vein patterns in the elderly population (aged 60 years and

above). A lack of inclusivity, has in the past ostracised senior citizens, from accessing basic amenities, such

as pension payments and healthcare services. A lack of inclusion of the elderly in ﬁnger vein biometric

research could result in the exclusion of elderly people from goods and services which use ﬁnger vein biometric

authentication. As the global population ages, ensuring the usability of biometric technologies for the elderly

is both a social and economic imperative.

1 INTRODUCTION

We live in a world where biometrics are commonly

used as an authentication method. From ﬁngerprint

authentication in our mobile devices to national iden-

tiﬁcation documents (Wagh et al., 2020), biometric

authentication is ingrained in the daily lives of many

people.

The ﬁnger vein biometric is a physiological bio-

metric that has been gaining popularity (Yang et al.,

2018). The intricate pattern of blood vessels pre-

sented underneath the epidermis of the skin, is what

we refer to as the vein pattern. This is a highly distinc-

tive pattern and can be used to uniquely identify indi-

viduals, including identical twins (Dev and Khanam,

2017).

Finger vein biometric capture is non-invasive,

making it ideal in the current pandemic climate.

Typically, the pattern is captured shining infra-red

light through a ﬁnger (Y

uksel et al., 2010). The

haemoglobin in the blood absorbs the light, mak-

ing the pattern appear darker and easier to capture

uksel et al., 2010). Veins lie under the skin, so

cannot be easily destroyed or manipulated (Dev and

Khanam, 2017). While these advantages make the ﬁn-

ger vein biometric ideal for identiﬁcation and authen-

tication, there is limited research on how ageing im-

https://orcid.org/0000-0003-4484-1920

https://orcid.org/0000-0002-9706-9420

pacts the need to re-enroll and how the current recog-

nition and capture methods perform with the elderly

population.

Age as a barrier in accessing biometric authen-

tication is not a new concept. The roll-out of the

Unique ID program in India saw the exclusion of the

elderly population due to difﬁculty of enrolment. Due

to lower elasticity of the skin in many older people,

ﬁngerprints were not able to be successfully captured.

The Unique ID is now used as a primary form of iden-

tiﬁcation across India, creating challenges for the el-

derly to access banking services and pension schemes

(Rebera and Guihen, 2012). There is an upward trend

in the percentage of the population above the age of

60 United Nations (2019), so elderly users need to be

considered when designing many technologies.

Recent years have seen the adoption of ﬁnger

vein recognition in commercial applications across

the globe. By 2011, 70% of the ﬁnancial institutions

in Japan were using ﬁnger vein biometric authentica-

tion systems for identiﬁcation and authentication of

customers (Wang et al., 2011). The ﬁnger vein au-

thentication device designed by Hitachi, is contact-

less and is used across most ATMs in Japan for au-

thentication.

Finger vein authentication systems have also been

successfully implemented in Poland (Hitachi, 2013).

The Polish bank, Bank BPH (Bank Przemysłowo-

Handlowy) has employed approximately 1800 ﬁn-

ger vein authentication systems across their branches.

472

Hall, J., John, J., Liebig, J. and Skariah, A.

Age Bias in Finger Vein Biometric Research.

DOI: 10.5220/0010876700003120

In Proceedings of the 8th International Conference on Information Systems Security and Privacy (ICISSP 2022), pages 472-477

ISBN: 978-989-758-553-1; ISSN: 2184-4356

The new technology replaces traditional means of au-

thentication: identity documents, passwords and PINs

(Hitachi, 2013). Now clients, need only declare their

identity, and are authenticated by the device.

Turkey uses ﬁnger vein authentication, with de-

vices implemented in hospitals and healthcare set-

tings (Hitachi Global, 2014b). The Hitachi ﬁnger vein

scanner has been implemented in the form of Wi-Fi

scanners to allow for convenient mobile authentica-

tion of patients. In addition to this, the system is used

as a means to manage payments and health insurance

beneﬁts (Hitachi Global, 2014b).

In the United Kingdom, the bank Barclays uses

ﬁnger vein recognition systems for customer authen-

tication. This technology was made available to their

corporate banking clients in 2015, to tackle iden-

tity fraud that was impacting UK businesses (Hitachi

Global, 2014a).

While advantageous, the widespread adoption of

ﬁnger vein recognition systems may prove to be a bur-

den for the elderly population. This paper explores

the research gap: the elderly population has not been

included in the implementation and study of ﬁnger

vein biometric authentication.

2 MEDICAL CONDITIONS THAT

IMPACT FINGER VEIN

BIOMETRIC

AUTHENTICATION

This section discusses the medical conditions that im-

pact the performance of ﬁnger vein recognition sys-

tems. The feature extraction method of ﬁnger vein

biometric systems are affected by vascular diseases

that manifest as a change in the pattern of the ﬁnger

vein. The elderly population are especially vulnera-

ble to vascular disease (Rodgers et al., 2019), which

may create an impact on the stability of a ﬁnger vein

biometric in the elderly population.

Diseases of the veins, such as vein thrombosis,

are caused by blood clots forming in the vein (CDC,

2020). This could affect the positioning and size of

the vein pattern used in ﬁnger vein recognition sys-

tems. If a blood clot is formed in the ﬁnger after a

person is enrolled, then they may fail at authentica-

tion, i.e., a false rejection. The elderly are particularly

susceptible to vein thrombosis, with almost 60% of

vein thrombosis incidents occurring in people above

the age of 70. The individual risk of incurring throm-

bosis is up to 15% for 90 year olds (Engbers et al.,

2010). When the events of vein thrombosis are com-

pared between the elderly and the younger popula-

tion, it becomes clear that aging is a prominent risk

factor for vein biometric authentication failure.

Connective tissue diseases, such as mixed con-

nective tissue disease (MCTD) or systemic sclerosis

(SSc), lead to the abnormal growth of connective tis-

sue (MedlinePlus, 2021). Diseases like SSc create

arterial and venous abnormalities, along with micro-

circulation, impacting the blood ﬂow, resulting in de-

fective circulation in ﬁngers (Allanore et al., 2007).

Reduction in blood ﬂow, conduces to a subsequent

reduction in available haemoglobin to absorb light,

impacting the performance of ﬁnger vein recognition

systems and increasing the possibility of false rejec-

tion (Kono et al., 2015). A study that compared the

recognition of ﬁnger vein authentication systems of

research participants suffering from connective tis-

sue disease against healthy research participants con-

ﬁrmed the higher rates of false rejection (Kono et al.,

2015). SSc usually appears in adults within the 40-50

age group. Late onset SSc manifests itself in the el-

derly population above the age of 60 (Rimar, 2020).

Thus, the elderly population are more vulnerable to

connective tissue diseases, and could be adversely im-

pacted due to the possibility of false rejection.

Injuries such as cuts and bruises can also nega-

tively impact ﬁnger vein recognition rates, as they

might be mistaken as veins (moﬁria Corporation,

2020). This is a cause of concern for the elderly, as

wounds tend to heal at slower rates when compared

to the younger population (University of Pittsburgh,

2020).

Ageing brings with it an increased risk of vascu-

lar diseases, connective tissue disorders coupled by

a slower rate of wound healing. Aging could have

an adverse impact on successful recognition rates in

ﬁnger vein authentication systems, resulting in false

rejection of the elderly trying to authenticate them-

selves.

Finger vein biometric authentication has been

proved to be useful, however the impact of the adop-

tion on the elderly population is yet to be investigated.

The world has seen a increase in the use of ﬁnger vein

biometric for authentication in banking and health-

care. A system that potentially excludes a population

demographic that comprises almost 20% of the adult

population in 2020 (United Nations, 2019). Any dif-

ﬁculty in authentication could potentially lock the el-

derly out of accessing ﬁnancial resources and critical

care.

The failure of the Unique ID program in India to

include the elderly population, emphasises the neces-

sity of assessing the performance of ﬁnger vein recog-

nition systems with elderly people.

Age Bias in Finger Vein Biometric Research

473

3 PUBLIC FINGER VEIN

DATABASES

There are several public ﬁnger vein databases that

have been used for the study of ﬁnger vein recogni-

tion systems. For a comprehensive evaluation of the

performance of any means of biometric authentica-

tion, the availability of datasets that reﬂect the user

demographic is a requirement. Making datasets avail-

able to the research community supports researchers

to ensure the ease of use and accuracy of the system.

However, most of the available datasets do not cap-

ture the ﬁnger vein biometrics of elderly people. This

section has conducted a investigation into the inclu-

sion of the elderly in publicly available ﬁnger vein

datasets.

3.1 Survey of Existing Literature on

Finger Vein Recognition

This paper investigates the existing research on ﬁnger

vein biometrics over the past decade. A major chal-

lenge that ﬁnger vein biometric research faces is the

unavailability of a large dataset. The lack of testing

in the elderly population has contributed a possible

bias against the senior population. The existing liter-

ature does not provide sufﬁcient evidence to guaran-

tee the applicability of their ﬁndings beyond the age

of 60. This is concerning, as recommendations are

made about the suitability of their applications in real

life scenarios, without taking into account the unavail-

ability of data on performance in the elderly popula-

tion (Liu and Song, 2012a). While there have been

assumptions, that the experiments conducted on sub-

jects below the age of 60 could model the behaviour

of vein patterns in the older population (Damak et al.,

2017), there is no concrete evidence that suggests age

based extrapolation is valid.

3.2 Methodology

We searched for publicly available databases and re-

search studies of ﬁnger vein biometrics. We limited

our research to papers that have been published in

English within the last decade. Anything published

before the year 2011 has not been considered in this

survey.

The publications and databases included in this

paper have been found by using the following key-

words in Google Scholar : “ﬁnger vein authentica-

tion”, “age bias in ﬁnger vein biometric research”,

“age bias in biometrics”, “ﬁnger vein recognition”.

Many of the databases do not have detailed in-

formation on the demographics of their research par-

ticipants. In these cases, world population statistics

(United Nations, 2019) are used to estimate the num-

ber of people above the age of 60.

• For the databases published between the years of

2011 and 2014, we use the population statistics of

2010, where people aged above 60 are 16% of the

adult population. Where no age detail is given, we

estimate that 16% of the research participants are

60 and above.

• For databases published between the years of

2015 and 2019, we use the population statistics of

2015, where people who are 60 and above com-

prise 17% of the adult population. Where no age

detail is given, we estimate that 17% of the re-

search participants are 60 and above.

3.3 Evaluation of Finger Vein Databases

Table 1 collates publicly available ﬁnger vein

databases and the different studies that have analysed

them. The total number of research participants for

each dataset is compared with the number of research

participants above the age of 60. Most datasets have

completely excluded the elderly population as shown

in the table.

Most datasets do not describe in detail the age dis-

tribution of the research participants. In those cases

an estimate was made, see section 3.2. Estimated

numbers are labeled with an asterisk (*). It is im-

portant to note that amongst the databases consid-

ered in this paper (see Table 1) , the databases MM-

CBNU 6000, PLUSVeinFV3 and PROTECT are the

only databases that state the inclusion of research par-

ticipants above the age of 60.

Overall, an estimated 6% of the total test popula-

tion are adults above the age of 60. Thus, compared

with 19% of the global adult population (United Na-

tions, 2019) the available data is insufﬁcient to sub-

stantiate that the elderly population are accounted for.

Extrapolating results from such a biased sample is

likely to draw biased conclusions.

4 CONCLUSIONS

In conclusion, very few people above the age of 60

have participated in ﬁnger vein research studies. The

longevity and stability of vein patterns cannot be

assumed. Medical conditions like vein thrombosis

and connective tissue diseases can impact the perfor-

mance of ﬁnger vein recognition systems, which can

lead to false rejection. The roll-out of ﬁnger vein sys-

tems for recognition and authentication could exclude

ICISSP 2022 - 8th International Conference on Information Systems Security and Privacy

474

Table 1: Public Finger Vein Databases, tabulated the number of research participants aged over 60 years.

* indicates estimated numbers.

S No Database name Total number

of research

participants

No of re-

search partici-

pants 60+

Published works that used the database Comments

1 MMCBNU 6000 100 3 (Lu et al., 2013), (Yang et al., 2014a),(Shaheed

et al., 2018)

2 SDUMLAHMT 106 0 (Yin et al., 2011), (Shaheed et al., 2018), (Yang

et al., 2014a), (Lu et al., 2018)

3 THUFVFDT 220 0* (Wenming et al., 2014), (Yang et al., 2014b),

(Yang et al., 2014a), (Shaheed et al., 2018)

The study was conducted in China, where the retirement age for is

60. The participants in the study were the staff and students of the

university, thus, none of the participants are above the age of 60.

4 HKPUFV 156 2* (Kumar and Zhou, 2011), (Yang et al., 2014a),

(Shaheed et al., 2018), (Lu et al., 2018)

93% of the participants were younger than 30 years. In 2010, adults

in the 60+ age bracket account for 16.2% of the adult world popula-

tion. (United Nations, 2019).

5 UTFV 60 2* (Ton and Veldhuis, 2013), (Yang et al., 2014a),

(Vanoni et al., 2014)

82% of the participants are in the age range of 19–30. In 2010, adults

in the 60+ age bracket account for 16.2% of the world adult popula-

tion (United Nations, 2019).

6 VERA 110 20* (Tome et al., 2015),(Vanoni et al., 2014),(Sha-

heed et al., 2018),(Tome et al., 2014)

The participants are aged between 18 and 60. The database was pub-

lished in 2015. In 2015, adults in the 60+ age bracket account for

17.71% of the adult world population (United Nations, 2019).

7 FVUSM 123 0 (Mohd Asaari et al., 2014), (Shaheed et al.,

2018)

8 PLUSVein-FV3 60 11* (Kauba et al., 2018) The participants are aged betwean 18-79 year. In 2015, adults in the

60+ age bracket account for 17.71% of the adult world population

(United Nations, 2019).

9 PROTECT multimodal

47 9* (University of Reading, 2017) The participants are aged between 21-76 years. In 2015, adults in the

60+ age bracket account for 17.71% of the adult world population

(United Nations, 2019).

10 KTDeaduk-FV 30 5* (Lee et al., 2011) In 2010, adults in the 60+ age bracket account for 16.2% of the adult

world population(United Nations, 2019).

11 CFVD 13 0 (Zhang et al., 2013)

12 S-EMB-Laser-FV 100 0 (Liu and Song, 2012b)

13 Gjøvik University Col-

lege, Norway -1

41 1* (Raghavendra et al., 2014) 90% of participants are between the ages of 20-35. In 2010, adults in

the 60+ age bracket account for 16.2% of the adult world population

(United Nations, 2019).

14 Wuhan University,

China

106 18* (Yang et al., 2012) The participants are between the ages of 19-60. In 2010, adults in

the 60+ age bracket account for 16.2% of the total world population

(United Nations, 2019).

15 Shandong. Univ 34 0 (Xi et al., 2013)

16 USM 51 0 (Rosdi et al., 2011)

17 Civil Aviation Univer-

sity of China

100 17* (Yang and Shi, 2014) In 2010, adults in the 60+ age bracket account for 16.2% of the adult

world population (United Nations, 2019).

18 Gjøvik University-2 125 0 (Raghavendra and Busch, 2015)

Total 1582 88*

Age Bias in Finger Vein Biometric Research

475

a large part of the dependant population from access-

ing basic amenities, including banking and health ser-

vices.

4.1 Recommendations for Future

Research

There is an upwards trend in the proportion of the

world’s population that are over the age of 60 (United

Nations, 2019). Thus, there is merit in observing

the way ﬁnger vein biometric authentication behaves

when implemented amongst the elderly. This includes

the ease of enrollment and the success rate of authen-

tication, including False Accept Rate (FAR) and False

Reject Rate (FRR). Therefore, research on ﬁnger vein

biometric systems needs to include elderly partici-

pants to ensure that ground-breaking technology does

not exclude a signiﬁcant portion of the population.

ACKNOWLEDGEMENTS

We would like to acknowledge the support of RMIT

University and CSIRO, that was crucial to the suc-

cessful completion of this paper.

REFERENCES

Allanore, Y., Seror, R., Chevrot, A., Kahan, A., and Drap

J. L. (2007). Hand vascular involvement assessed by

magnetic resonance angiography in systemic sclero-

sis. Arthritis and rheumatism, 56(8):2747–2754.

CDC (2020). What is venous thromboembolism? https:

//www.cdc.gov/ncbddd/dvt/facts.html.

Damak, W., Boukhris Trabelsi, R., Damak Masmoudi, A.,

Sellami, D., and Nait-Ali, A. (2017). Age and gender

classiﬁcation from ﬁnger vein patterns. In Madureira,

A. M., Abraham, A., Gamboa, D., and Novais, P.,

editors, Intelligent Systems Design and Applications,

pages 811–820, Cham. Springer International Pub-

lishing.

Dev, R. and Khanam, R. (2017). Review on ﬁnger vein fea-

ture extraction methods. In 2017 International Con-

ference on Computing, Communication and Automa-

tion (ICCCA), pages 1209–1213.

Engbers, M. J., van Hylckama Vlieg, A., and Rosendaal,

F. R. (2010). Venous thrombosis in the elderly: inci-

dence, risk factors and risk groups: Risk factors for

venous thrombosis in the elderly population. Journal

of thrombosis and haemostasis, 8(10):2105–2112.

Hitachi (2013). Finger vein technology for bank

bph (poland). https://www.hitachi.eu/en/case-studies/

ﬁnger-vein-technology-bank-bph-poland.

Hitachi Global (2014a). Barclays ﬁrst in uk to launch new

biometric reader for customers. https://www.hitachi.

com/New/cnews/month/2014/09/140905.html.

Hitachi Global (2014b). Hitachi and mig introduce

mobile wi-ﬁ biometric scanners in turkish hos-

pitals. https://www.hitachi.com/New/cnews/month/

2014/07/140709a.html.

Kauba, C., Prommegger, B., and Uhl, A. (2018). Fo-

cussing the beam - a new laser illumination based data

set providing insights to ﬁnger-vein recognition. In

2018 IEEE 9th International Conference on Biomet-

rics Theory, Applications and Systems (BTAS), pages

1–9.

Kono, M., Miura, N., Fujii, T., Ohmura, K., Yoshifuji,

H., Yukawa, N., Imura, Y., Nakashima, R., Ikeda, T.,

Umemura, S., Miyatake, T., and Mimori, T. (2015).

Personal authentication analysis using ﬁnger-vein pat-

terns in patients with connective tissue diseases - pos-

sible association with vascular disease and seasonal

change. PloS one, 10(12):e0144952–e0144952.

Kumar, A. and Zhou, Y. (2011). Human identiﬁcation using

ﬁnger images. IEEE transactions on image processing

: a publication of the IEEE Signal Processing Society,

21:2228–44.

Lee, E. C., Jung, H., and Kim, D. (2011). New ﬁnger bio-

metric method using near infrared imaging. Sensors,

11(3):2319–2333.

Liu, Z. and Song, S. (2012a). An embedded real-

time ﬁnger-vein recognition system for mobile de-

vices. IEEE transactions on consumer electronics,

58(2):522–527.

Liu, Z. and Song, S. (2012b). An embedded real-

time ﬁnger-vein recognition system for mobile de-

vices. IEEE Transactions on Consumer Electronics,

58(2):522–527. Copyright - Copyright The Institute

of Electrical and Electronics Engineers, Inc. (IEEE)

May 2012; Last updated - 2012-07-07.

Lu, Y., Xie, S. J., Yoon, S., Wang, Z., and Park, D. S.

(2013). An available database for the research of

ﬁnger vein recognition. In 2013 6th International

Congress on Image and Signal Processing (CISP),

volume 01, pages 410–415.

Lu, Y., Yang, G., Yin, Y., and Xi, X. (2018). Finger vein

recognition with anatomy structure analysis. IEEE

Transactions on Circuits and Systems for Video Tech-

nology, 28(8):1892–1905. Copyright - Copyright The

Institute of Electrical and Electronics Engineers, Inc.

(IEEE) 2018; Last updated - 2019-02-01.

MedlinePlus (2021). Scleroderma. https://medlineplus.gov/

scleroderma.html.

moﬁria Corporation (2020). Frequently asked questions.

https://www.moﬁria.com/en/faq.

Mohd Asaari, M. S., Suandi, S. A., and Rosdi, B. A.

(2014). Fusion of band limited phase only correla-

tion and width centroid contour distance for ﬁnger

based biometrics. Expert systems with applications,

41(7):3367–3382.

Raghavendra, R. and Busch, C. (2015). Exploring dorsal

ﬁnger vein pattern for robust person recognition. In

2015 International Conference on Biometrics (ICB),

pages 341–348.

ICISSP 2022 - 8th International Conference on Information Systems Security and Privacy

476

Raghavendra, R., Raja, K. B., Surbiryala, J., and Busch, C.

(2014). A low-cost multimodal biometric sensor to

capture ﬁnger vein and ﬁngerprint. In IEEE Interna-

tional Joint Conference on Biometrics, pages 1–7.

Rebera, A. P. and Guihen, B. (2012). Biometrics for an age-

ing society societal and ethical factors in biometrics

and ageing. In 2012 BIOSIG - Proceedings of the In-

ternational Conference of Biometrics Special Interest

Group (BIOSIG), pages 1–4.

Rimar, D. (2020). Systemic Sclerosis in the Elderly, pages

207–228. Springer International Publishing, Cham.

Rodgers, J. L., Jones, J., Bolleddu, S. I., Vanthenapalli, S.,

Rodgers, L. E., Shah, K., Karia, K., and Panguluri,

S. K. (2019). Cardiovascular risks associated with

gender and aging. Journal of cardiovascular devel-

opment and disease, 6(2):19.

Rosdi, B. A., Shing, C. W., and Suandi, S. A. (2011). Finger

vein recognition using local line binary pattern. Sen-

sors, 11(12):11357–11371.

Shaheed, K., Liu, H., Yang, G., Qureshi, I., Gou, J., and

Yin, Y. (2018). A systematic review of ﬁnger vein

recognition techniques. Information, 9(9).

Tome, P., Raghavendra, R., Busch, C., Tirunagari, S., Poh,

N., Shekar, B. H., Gragnaniello, D., Sansone, C., Ver-

doliva, L., and Marcel, S. (2015). The 1st competition

on counter measures to ﬁnger vein spooﬁng attacks. In

2015 International Conference on Biometrics (ICB),

pages 513–518.

Tome, P., Vanoni, M., and Marcel, S. (2014). On the vul-

nerability of ﬁnger vein recognition to spooﬁng. In

2014 International Conference of the Biometrics Spe-

cial Interest Group (BIOSIG), pages 1–10.

Ton, B. T. and Veldhuis, R. N. J. (2013). A high quality

ﬁnger vascular pattern dataset collected using a cus-

tom designed capturing device. In 2013 International

Conference on Biometrics (ICB), pages 1–5.

United Nations (2019). World population prospects 2019.

https://population.un.org/wpp/Download/Standard/

Population/.

University of Pittsburgh (2020). Why do

older people heal more slowly? https:

//www.pitt.edu/pittwire/features-articles/

why-do-older-people-heal-more-slowly.

University of Reading (2017). Protect multimodal dataset.

http://projectprotect.eu/dataset/.

Vanoni, M., Tome, P., El Shafey, L., and Marcel, S. (2014).

Cross-database evaluation using an open ﬁnger vein

sensor. In 2014 IEEE Workshop on Biometric Mea-

surements and Systems for Security and Medical Ap-

plications (BIOMS) Proceedings, pages 30–35.

Wagh, D. P., Fadewar, H. S., and Shinde, G. N. (2020). Bio-

metric ﬁnger vein recognition methods for authentica-

tion. In Iyer, B., Deshpande, P. S., Sharma, S. C., and

Shiurkar, U., editors, Computing in Engineering and

Technology, pages 45–53, Singapore. Springer Singa-

pore.

Wang, K., Ma, H., Popoola, O., and Liu, J. (2011). Finger

vein recognition.

Wenming, Y., Huang, X., Zhou, F., and Liao, Q. (2014).

Fusion of ﬁnger vein and ﬁnger dorsal texture for per-

sonal identiﬁcation based on comparative competitive

coding. Information Sciences, 268:20–32.

Xi, X., Yang, G., Yin, Y., and Meng, X. (2013). Finger

vein recognition with personalized feature selection.

Sensors, 13(9):11243–11259.

Yang, G., Xi, X., and Yin, Y. (2012). Finger vein recog-

nition based on a personalized best bit map. Sensors,

12(2):1738–1757.

Yang, J. and Shi, Y. (2014). Towards ﬁnger-vein image

restoration and enhancement for ﬁnger-vein recogni-

tion. Information sciences, 268:33–52.

Yang, L., Yang, G., Yin, Y., and Xi, X. (2018). Finger vein

recognition with anatomy structure analysis. IEEE

Transactions on Circuits and Systems for Video Tech-

nology, 28(8):1892–1905.

Yang, L., Yang, G., Yin, Y., and Zhou, L. (2014a). A survey

of ﬁnger vein recognition. In Sun, Z., Shan, S., Sang,

H., Zhou, J., Wang, Y., and Yuan, W., editors, Bio-

metric Recognition, pages 234–243, Cham. Springer

International Publishing.

Yang, W., Huang, X., Zhou, F., and Liao, Q. (2014b). Com-

parative competitive coding for personal identiﬁcation

by using ﬁnger vein and ﬁnger dorsal texture fusion.

Information sciences, 268:20–32.

Yin, Y., Liu, L., and Sun, X. (2011). Sdumla-hmt: A mul-

timodal biometric database. In Sun, Z., Lai, J., Chen,

X., and Tan, T., editors, Biometric Recognition, pages

260–268, Berlin, Heidelberg. Springer Berlin Heidel-

berg.

uksel, A., Akarun, L., and Sankur, B. (2010). Biomet-

ric identiﬁcation through hand vein patterns. In 2010

International Workshop on Emerging Techniques and

Challenges for Hand-Based Biometrics, pages 1–6.

Zhang, C., Li, X., Liu, Z., Zhao, Q., Xu, H., and Su, F.

(2013). The cfvd reﬂection-type ﬁnger-vein image

database with evaluation baseline. In Sun, Z., Shan,

S., Yang, G., Zhou, J., Wang, Y., and Yin, Y., edi-

tors, Biometric Recognition, pages 282–287, Cham.

Springer International Publishing.

Age Bias in Finger Vein Biometric Research

477