A Methodology for Biometric Identification Using Vein ...
Transcript of A Methodology for Biometric Identification Using Vein ...
International Journal of Contemporary Mathematical Sciences
Vol. 16, 2021, no. 4, 135 – 148
HIKARI Ltd, www.m-hikari.com
https://doi.org/10.12988/ijcms.2021.91610
A Methodology for Biometric Identification Using
Vein Geometry in Infrared Images
of the Back of the Hand
B. Luna-Benoso
Instituto Politécnico Nacional. Escuela Superior de Cómputo. Av. Juan de Dios Batíz,
esq. Miguel Othón de Mendizábal, México City 07738, México
J. C. Martínez-Perales
Instituto Politécnico Nacional. Escuela Superior de Cómputo. Av. Juan de Dios Batíz,
esq. Miguel Othón de Mendizábal, México City 07738, México
J. Cortés-Galicia
Instituto Politécnico Nacional. Escuela Superior de Cómputo. Av. Juan de Dios Batíz,
esq. Miguel Othón de Mendizábal, México City 07738, México
This article is distributed under the Creative Commons by-nc-nd Attribution License.
Copyright © 2021 Hikari Ltd.
Abstract
With the great boom in technological applications of aspects such as computer security
and access control among others, the need arises to create methodologies capable of
identifying individuals. Identification through biometric features using infrared images
is considered one of the safest modalities within biometrics. In this work, a
methodology is proposed that allows biometric identification from infrared images of
the back of the hand. The methodology consists of 4 modules: 1. Image acquisition, 2.
Image segmentation, 3. Characteristics extraction and 4. Classification. Image
acquisition was carried out through the proposal of an electronic prototype that allows
the capture of infrared images, subsequently the images are segmented to obtain the
136 B. Luna-Benoso, J. C. Martínez-Perales and J. Cortés-Galicia
geometry of the veins by skeletonizing the vascular network of the back of the hand,
to This, methods in the spatial domain and a set of 11 geometric transformations are
used from which characteristics are extracted using Hu moments, finally, for the
classification, the Random Forest model was used. The proposed methodology was
compared with other models via the results obtained by the WEKA platform. The
model yielded an accuracy of 90.20%.
Keywords: Pattern recognition, image analysis, biometrics, random forest.
1. Introduction
The word biometry derives from the Greek bio (life) and metry (measure), from
which the term biometry is inferred as the measure of life [1]. Since its inception, the
field of biometrics has focused on measuring the distinctive features of a person as a
way to authenticate them [2]. The term biometrics is used to encompass different
technologies applied to the identification and authentication of people through their
physical, chemical, behavioral or distinctive features [3]. With the great boom in
technological applications, the need arises to propose methodologies that allow
safeguarding the security, confidentiality and integrity of people's information applied
to electronic commerce and bank transfers, among others [4, 5], in this case, systems
Biometrics are an option for computer security.
There are different biometric systems that make use of different biometric
approaches, some of these approaches are fingerprint recognition [6, 7], iris [8, 9],
facial recognition [10, 11], recognition by signature [12, 13], voice recognition [14,
15], hand geometry [16, 17] and those based on vein geometry [18, 19] among others.
Biometric systems that make use of the geometry of the veins do so using the geometry
of the veins in the fingers [20, 21], the geometry of the veins in the forearm [22, 23],
the geometry of the veins of the palm of the hand [24, 25], or use the geometry of the
veins on the back of the hands [18, 26]. Some works make use of thermal images [27,
28] and others of infrared images [20, 29] for the biometric recognition of the geometry
of the veins.
The main advantages of biometric systems based on the geometry of the veins using
infrared images are that they are non-invasive, they are very precise, difficult to
reproduce, they are non-intrusive, and they are also distinctive and unique to each
individual [30]. On the other hand, infrared light with wavelengths between 740 nm to
940 nm can penetrate the skin tissue up to 5mm in depth, reaching the deoxygenated
hemoglobin, thus forming a dark contrast with the skin tissue [31] showing the veins
as dark lines, from which the geometry of the veins is obtained. Various methodologies
have been developed that allow the identification of people through infrared images of
the veins on the back of the hands, some of these works are as follows. Pal et al. [32]
propose a methodology in which the images are obtained by means of an IR Webcam
Methodology for biometric identification using vein geometry 137
to which an equalization of the histogram and the Hough transform are carried out to
locate the veins in the hand, the extraction of characteristics is done by means of minute
points that include bifurcations and end points, finally to measure the similarity
between two images that allow identification, the Hausdorff distance is used. Park et
al. [33] capture infrared images using a camera equipped with an infrared light emitting
diode, IR filter, mirror and hand support, the image is decomposed into three sub-
images (side view of the hand, back of the hand and vascular pattern) , the extraction
of characteristics includes the thickness of the lateral view of the hand, angles of the
valleys of the fingers, the curvature K and the vascular pattern, the similarity for the
identification is carried out using the Euclidean distance, the distance measured by the
polygon curves and a matching algorithm. Ferrer et al. [34] capture infrared images of
the back of the hand using an array of 21 LEDs with a wavelength of 940nm and a
CCD camera with an infrared filter, then obtain a template of the vein pattern applying
mathematical morphology and Gabor filters, for the identification make use of
similarity by means of Hamming distance. Chen et al. [35] propose a methodology in
which the captured images are made using a CCD camera and an 850 nm infrared LED
matrix. The authors study the difference between signals near prostheses and on the
living body, for identification they combine the method correction, rotation and
translation in conjunction with the PLBP and SVM algorithm. On the other hand, Khan
et al. [36] analyze the factors that affect the biometry of the hand during the capture of
the images, such as the orientation angle of the hand, the distance between the camera
and the light, the color of the skin and the temperature, among others.
In this work, a methodology is proposed that allows the biometric identification of
an individual through the geometry of the veins in infrared images of the back of the
hand. This work is structured as follows. Section 1 shows the introduction, section 2
shows the methodology, section 3 the experiments and results carried out, and section
4 the conclusions.
2. Methodology
Figure 1 (p. 137 below) shows the general methodology used to identify people
through the geometry of the back of the hand. Later each of the modules that compose
it are shown.
2.1.Image Acquisition
The acquisition of the images is the first module that makes up the proposed
methodology, for this, a wooden box measuring 24cm × 22cm × 14cm was built in
black, without the front of the box where the hand is inserted, with a removable top cap
that contains a 3cm radius circumference in the center to allow the passage of the lens
of a 7.2 MP digital camera type Canon PowerShot A470 to which an infrared filter is
attached. On the back face of the box there is also a 3cm circumference where a vein
locator device is secured with infrared LEDs with a wavelength of 850 nm in the shape
138 B. Luna-Benoso, J. C. Martínez-Perales and J. Cortés-Galicia
of a cylindrical bar where the user places the hand pressing and it emits the Infrared
light through the palm of the hand towards the back, revealing the vascular network on
the back of the hand. Figure 2 shows the image of the prototype that allows the capture
of the images.
Fig. 1. Methodology used for biometric identification using the back of the hand.
Fig. 2. Prototype that allows capturing infrared images of the back of the hand
Methodology for biometric identification using vein geometry 139
2.2.Segmentation
A series of passes were carried out that allow the infrared image to be improved
from the beginning until the skeletonization of the geometry of the veins on the back
of the hand was obtained. The steps performed are listed below.
1. Media filter. A 3 × 3 masked media filter is applied to soften the edges or
textures of the image.
2. Weighted average filter. Continuing with the goal of smoothing the image, a
weighted median filter with a 3 × 3 mask is applied with a center value of 2.
3. Cleared the image. A logarithmic stretch is applied to lighten the image such
that pixels with a dark hue are lightened and those with a light hue remain.
4. Grayscale. The grayscale image is obtained by the formula
𝐼(𝑥, 𝑦) = 0.299 ∗ 𝑅(𝑥, 𝑦) + 0.587 ∗ 𝐺(𝑥, 𝑦) + 0.114 ∗ 𝐵(𝑥, 𝑦) where 𝑅(𝑥, 𝑦), 𝐺(𝑥, 𝑦) y 𝐵(𝑥, 𝑦) represent the value of the red, green and blue
plane on the pixel with coordinates (𝑥, 𝑦). 5. Lightened the grayscale image. Logarithmic expansion is applied again to
lighten the grayscale image.
6. Thresholding using Niblack. The Niblack thresholding method is a local type
method, this method is tolerant to changes in lighting. Niblack proposes an
algorithm that obtains local thresholds from a sliding mask 𝑤(𝑥, 𝑦), to obtain
the mean 𝜇(𝑥, 𝑦) and the variance 𝜎(𝑥, 𝑦). The threshold value 𝑇(𝑥, 𝑦) is given
by 𝑇(𝑥, 𝑦) = 𝜇(𝑥, 𝑦) + 𝑘 ∙ 𝜎(𝑥, 𝑦). 7. Morphological Lock and Opening. The morphological operation of the lock is
applied, followed by the morphological operation of the opening with a
structuring element equal to the Moore neighborhood.
8. Image negative. The image is inverted, that is, the white pixels become black
and the black ones become white through the function 𝑁𝑒𝑔(𝑥, 𝑦) = 255 −𝐼(𝑥, 𝑦).
9. Median filter. A median filter is applied to reduce blurring from having
previously applied a median filter.
10. Related components. All components connected through 8-connectivity are
tagged and those with an area smaller than 200 × 200 pixels are removed.
11. Image negative. The image is reversed again.
12. Morphological lock. A morphological lock operation is used with a diamond-
shaped structuring element with a radius of 10 pixels.
13. Morphological dilation. A morphological dilation operation is applied with a
structuring element in the form of a vertical line with a radius of 10 pixels.
14. Morphological lock. A morphological lock operation is applied with a
structuring element in the shape of a circle with a radius of 5 pixels.
15. Morphological skeletonization. The morphological skeletonization process is
applied that allows us to obtain the vascular network of the back of the hand.
140 B. Luna-Benoso, J. C. Martínez-Perales and J. Cortés-Galicia
Figure 3 shows the application of steps 1 to 15 that make it possible to obtain the
skeletonization of the geometry of the veins of the back of the hand applied to an
example of an infrared image of the back of the hand.
Fig. 3. Segmentation process of the vascular network on the back of the hand. In (a)
original image, after the result of applying in (b) step 1 and 2, in (c) step 3, (d) step 4,
(e) step 5, in (f) step 6, in (g) step 7, at (h) step 9, at (i) step 10, at (j) step 11, at (k) step
12, at (l) step 13, at (m) step 14, at (n) step 15 and finally in (o) the skeletonization of
the geometry of the veins of the back of the hand is obtained.
2.3.Feature extraction
Once the skeletonization of the geometry of the veins on the back of the hand has
been obtained in infrared images as the result shown in figure 3, a total of 11
transformations are applied to the result of the skeletonized image, these
transformations correspond to 45 °, 90 °, 135 °, 180 °, 225 °, 270 ° and 315 ° rotations;
a zoom of 0.85 and one of 1.15; a horizontal reflection and a vertical reflection (figure
Methodology for biometric identification using vein geometry 141
4). From each of these images obtained, the 7 Hu moments that are shown below are
extracted.
Φ1 = 𝜂20 + 𝜂02
Φ2 = (𝜂20 − 𝜂02)2 + 4𝜂11
2 Φ3 = (𝜂30 − 3𝜂012)
2 + (3𝜂21 − 𝜂03)2
Φ4 = (𝜂03 + 𝜂12)2 + (𝜂21 + 𝜂03)
2 Φ5 = (𝜂30 − 3𝜂12)(𝜂30 + 𝜂12)[(𝜂30 + 𝜂12)
2 − 3(𝜂21 + 𝜂03)2]
+ (3𝜂21 − 𝜂03)(𝜂21 + 𝜂03)[3(𝜂30 + 𝜂12)2 − (𝜂21 + 𝜂03)
2] Φ6 = (𝜂20 − 𝜂02)[(𝜂30 + 𝜂12)
2 − (𝜂21 + 𝜂03)2] + 4𝜂11(𝜂30 + 𝜂12)(𝜂21 + 𝜂03)
Φ7 = (3𝜂21 − 𝜂03)(𝜂30 + 𝜂12)[(𝜂30 + 𝜂12)2 − 3(𝜂21 + 𝜂03)
2]− (𝜂30 − 3𝜂12)(𝜂21 + 𝜂03)[3(𝜂30 + 𝜂12)
2 − (𝜂21 + 𝜂03)2]
where 𝜂𝑟𝑠 = 𝜇𝑟𝑠 𝜇00⁄ , 𝑡 = (𝑟 + 𝑠) 2⁄ + 1, 𝜇𝑟𝑠 = ∑ (𝑖 − 𝑖)̅𝑟(𝑗 − 𝑗)̅𝑠𝑖,𝑗∈𝑅𝐸𝐺 , 𝑖̅ = 𝑚10 𝑚00⁄ , 𝑗̅ =
𝑚01 𝑚00⁄ , 𝑚𝑟𝑠 = ∑ 𝑖𝑟𝑗𝑠𝑖,𝑗∈𝑅𝐸𝐺 with 𝑟, 𝑠 ∈ ℕ and 𝑅𝐸𝐺 the set of pixels within the region.
Fig 4. In (a) original image, in (b) 45 ° rotation, in (c) 90 ° rotation, in (d) 135 °
rotation, in (e) 180 ° rotation, in (f) 225 ° rotation, in (g) 270 ° rotation, in (h) rotation
of 315, in (i) zoom of 1.15, in (j) zoom of 0.85, in (k) vertical reflection and in (l)
reflection horizontal.
2.4.Clasification
For the classification, use is made of Random Forest whose output is obtained from
a set of decision trees and the decision with the highest votes is the output of the
algorithm. Figure 5 shows the general scheme of the Random Forest algorithm.
142 B. Luna-Benoso, J. C. Martínez-Perales and J. Cortés-Galicia
Fig. 5. Random Forest Classifier.
3. Experimentation and results
To carry out the experiments, a total of 20 people were considered who were
captured 10 different shots of the back of the hand using the prototype exposed in
section 2.1. The methodology shown in section 2.2 was applied to each of the infrared
images obtained by the users to obtain the skeletonization of the geometry of the back
of the hand, to the result of this skeletonization, the 11 transformations exposed in the
section 2.3 obtaining a total of 2400 images of the skeleton of the geometry of the back
of the hand. To carry out the classification, 80% were used to make up the training set
and the remaining 20% to make up the test set, that is, 1920 images were used to make
up the training set and 480 to perform the tests. The Random Forest algorithm was
implemented, and the WEKA platform was used to make the comparison with other
classifiers. WEKA is a free software platform designed to be distributed under the
GNU-GPL license and used to test machine learning and data mining. WEKA was
written in Java and developed by the University of Waikato, New Zealand. The
algorithms considered to perform the comparison were SVM, MLP, kNN and ID3
decision trees. Table 1 shows the results of the accuracy of each of these algorithms
and is compared with Random Forest.
Methodology for biometric identification using vein geometry 143
Classifier Accuracy
SVM 82.91%
MLP 84.37%
kNN 84.37%
ID3 84.37%
Random Forest 90.20%
Table 1. Accuracy shown by various classifiers using the WEKA platform
As can be seen in table 1, with the proposed methodology that allows to obtain the
skeletonization of the geometry of the veins of the back of the hand in infrared images
in conjunction with the proposed transformations from which the Hu moments are
extracted and making use of the classifier Random Forest, the highest accuracy of
90.20% is obtained.
4. Conclusions
The present work shows the proposal of a methodology that allows biometric
identification through infrared images of the back of the hand from which the pattern
of vein geometry is extracted by skeletonization using different algorithms based on
the spatial domain. Once the geometry of the back of the hand has been extracted, it is
proposed to use 11 transformations that are applied to the skeletonized images. These
11 transformations are 45 °, 90 °, 135 °, 180 °, 225 °, 270 °, and 315 ° rotations; a zoom
of 0.85 and one of 1.15; a horizontal reflection and a vertical reflection. The Hu
moments that make up a vector of characteristics of a person for each transformation
are calculated for each of these images. A total of 2400 images of 20 different people
were used for the experiments. The results of the accuracy show that the best
performance is the Random Forest algorithm with an accuracy of 90.20%. In this way,
with the proposal of the methodology from the acquisition of the images, segmentation
and extraction of characteristics, competitive results are obtained using the Random
Forest classifier.
Acknowledgments. The authors would like to thank the Instituto Politécnico Nacional
(Secretaría Académica, COFAA, EDD, SIP and ESCOM) for their economical support
to develop this work.
144 B. Luna-Benoso, J. C. Martínez-Perales and J. Cortés-Galicia
References
[1] Chauhan, S., Arora, A. S., & Kaul, A., A survey of emerging biometric modalities.
Procedia Computer Science, 2 (2010), 213-218.
https://doi.org/:10.1016/j.procs.2010.11.027
[2] Dantcheva, A., Velardo, C., D’Angelo, A., & Dugelay, J.-L., Bag of soft biometrics
for person identification. Multimedia Tools and Applications, 51 (2) (2010), 739–
77. https://doi.org/:10.1007/s11042-010-0635-7
[3] Van den Broek, E. L., Beyond biometrics. Procedia Computer Science, 1 (1) (2010),
2511–2519. https://doi.org/:10.1016/j.procs.2010.04.284
[4] Miller, D., & Fauve, B., Mobile e-commerce to drive voice-based authentication.
Biometric Technology Today, 2012 (2012) (2), 5–8. https://doi.org/:10.1016/s0969-
4765(12)70053-x
[5] Dobbie, S., Challenge of biometric security for banks. Biometric Technology
Today, 2020 (2020) (3), 5–7. https://doi.org/:10.1016/s0969-4765(20)30037-0
[6] Joshi, M., Mazumdar, B., & Dey, S., A Comprehensive Security Analysis of Match-
in-database Fingerprint Biometric System. Pattern Recognition Letters,
(2020). https://doi.org/:10.1016/j.patrec.2020.07.024
[7] Sharma, A., Arya, S., & Chaturvedi, P., A Novel Image Compression Based Method
for Multispectral Fingerprint Biometric System. Procedia Computer Science, 171
(2020), 1698–1707. https://doi.org/:10.1016/j.procs.2020.04.182
[8] Murad, M., Key markets for iris biometrics, Biometric Technology Today, 2021,
Issue 7-8 (2021), 5-7. https://doi.org/.org/10.1016/S0969-4765(21)00082-5
[9] Kang, J.-S., Mobile iris recognition systems: An emerging biometric technology.
Procedia Computer Science, 1 (1) (2010), 475–484.
https://doi.org/:10.1016/j.procs.2010.04.051
Methodology for biometric identification using vein geometry 145
[10] Tome, P., Vera-Rodriguez, R., Fierrez, J., & Ortega-Garcia, J., Facial soft
biometric features for forensic face recognition. Forensic Science International, 257
(2015), 271–284. https://doi.org/:10.1016/j.forsciint.2015.09.002
[11] Tiong, L. C. O., Kim, S. T., & Ro, Y. M., Multimodal facial biometrics
recognition: Dual-stream convolutional neural networks with multi-feature fusion
layers. Image and Vision Computing, (2020), 103977.
https://doi.org/:10.1016/j.imavis.2020.103977
[12] Maiorana, E., Biometric cryptosystem using function based on-line signature
recognition. Expert Systems with Applications, 37 (4) (2010), 3454–3461.
https://doi.org/:10.1016/j.eswa.2009.10.043
[13] Bailador, G., Sanchez-Avila, C., Guerra-Casanova, J., & de Santos Sierra, A.,
Analysis of pattern recognition techniques for in-air signature biometrics. Pattern
Recognition, 44 (10-11) (2011), 2468–2478.
https://doi.org/:10.1016/j.patcog.2011.04.010
[14] Abdalrahman, R. S. A., Bolat, B., & Kahraman, N., A cascaded voice biometric
system, Procedia Computer Science, 131 (2018), 1223–1228.
https://doi.org/:10.1016/j.procs.2018.04.334
[15] Kuśmierczyk A., Sławińska M., Żaba K., Saeed K., Biometric Fusion System
Using Face and Voice Recognition. In: Chaki R., Cortesi A., Saeed K., Chaki N. (eds.)
Advanced Computing and Systems for Security, Advances in Intelligent Systems and
Computing, vol 996. Springer, Singapore, 2020.
https:// https://doi.org/.org/10.1007/978-981-13-8969-6_5
[16] Iula, A., Hine, G., Ramalli, A., & Guidi, F., An Improved Ultrasound System for
Biometric Recognition Based on Hand Geometry and Palmprint. Procedia
Engineering, 87 (2014), 1338–1341. https://doi.org/:10.1016/j.proeng.2014.11.709
[17] Mohammed, H., H., Baker, S., Nori, A., Biometric identity Authentication System
Using Hand Geomtry Measurements, Journal of Physics: Conference Series.
ICMAICT 2020, IOP Publishing, 1804 (2021) 012144. https://doi.org/:10.1088/1742-
6596/1804/1/012144
146 B. Luna-Benoso, J. C. Martínez-Perales and J. Cortés-Galicia
[18] Jia, W., Xia, W., Zhang, B., Zhao, Y., Fei, L., Kang, W., Huang, D., Guo, G., A
survey on dorsal hand vein biometrics, Pattern Recognition, 120 (2021) 108122.
https:// https://doi.org/.org/10.1016/j.patcog.2021.108122
[19] Wagh D.P., Fadewar H.S., Shinde G.N., Biometric Finger Vein Recognition
Methods for Authentication. In: Iyer B., Deshpande P., Sharma S., Shiurkar U. (eds)
Computing in Engineering and Technology. Advances in Intelligent Systems and
Computing, vol 1025, 2020. Springer, Singapore.
https:// https://doi.org/.org/10.1007/978-981-32-9515-5_5
[20] Ma, H., Hu, N., Fang, C., The biometric recognition system base don near-infrared
finger vein image, Infrared Physics and Technology, 116 (2021) 103734.
https:// https://doi.org/.org/10.1016/j.infrared.2021.103734
[21] Hu, N., Ma, H., & Zhan, T., Finger vein biometric verification using Block Multi-
scale Uniform Local Binary Pattern features and Block Two-Directional Two-
Dimension Principal Component Analysis, Optik, (2020), 163664.
https://doi.org/:10.1016/j.ijleo.2019.163664
[22] Lui, H., Tang, C., Li, Z., Zhang, Y., Wang, B., Multimodal biometric method
based on the score-level fusion of forearm vein and forearm geometry. ICVISP 2020:
Proceedings of the 2020 4th International Conference on Vision, Image and Signal
Processing, No 41 (2020), 1-6. https://doi.org/.org/10.1145/3448823.3448840
[23] Choraś R.S., Biometric Identification Based on Forearm Vein Pattern. In: Tavares
J., Natal Jorge R. (eds.) VipIMAGE 2019. VipIMAGE 2019. Lecture Notes in
Computational Vision and Biomechanics, vol 34, 2019, Springer, Cham.
https:// https://doi.org/.org/10.1007/978-3-030-32040-9_6
[24] Adiraju, R. V., Masanipalli, K. K., Reddy, T. D., Pedapalli, R., Chundru, S., &
Panigrahy, A. K., An extensive survey on finger and palm vein recognition system.
Materials Today: Proceedings (2020). https://doi.org/:10.1016/j.matpr.2020.08.742
[25] Athale, S. S., Patil, D., Deshpande, P., & Dandawate, Y. H., Hardware
Implementation of Palm Vein Biometric Modality for Access Control in Multilayered
Methodology for biometric identification using vein geometry 147
Security System, Procedia Computer Science, 58 (2015), 492–498.
https://doi.org/:10.1016/j.procs.2015.08.013
[26] Park, G., & Kim, S., Hand Biometric Recognition Based on Fused Hand Geometry
and Vascular Patterns, Sensors, 13 (3) (2013), 2895–2910.
https://doi.org/:10.3390/s130302895
[27] Kumar, A., Hanmandlu, M., Madasu, V. K., & Lovell, B. C., Biometric
Authentication Based on Infrared Thermal Hand Vein Patterns. 2009 Digital Image
Computing: Techniques and Applications, 2019.
https://doi.org/:10.1109/dicta.2009.63
[28] Alpar, O., A Novel Fuzzy Curvature Method for Recognition of Anterior Forearm
Subcutaneous Veins by Thermal Imaging, Expert Systems with Applications, 2018.
https://doi.org/:10.1016/j.eswa.2018.11.017
[29] Gupta, P., Srivastava, S., & Gupta, P., An accurate infrared hand geometry and
vein pattern based authentication system, Knowledge-Based Systems, 103 (2016), 143–
155. https://doi.org/:10.1016/j.knosys.2016.04.008
[30] T. Sabhanayagam, V. Prasanna Venkatesan and K. Senthamaraikannan, A
Comprehensive Survey on Various Biometric Systems, International Journal of
Applied Engineering Research, 13 (5) (2018), 2276-2297.
[31] Cheng-Tang Pan, Mark D. Francisco, Chung-Kun Yen,Shao-Yu Wang, Yow-
Ling Shiue. Vein Pattern Locating Technology for Cannulation: A Review of the Low-
Cost Vein Finder Prototypes Utilizing near Infrared (NIR) Light to Improve Peripheral
Subcutaneous Vein Selection for Phlebotomy, 19 (16) (2019), 3573.
[32] Pal, M. M., & Jasutkar, R. W., Implementation of Hand Vein Structure
Authentication Based System, 2012 International Conference on Communication
Systems and Network Technologies, 2012. https://doi.org/:10.1109/csnt.2012.34
[33] Park, G., & Kim, S., Hand Biometric Recognition Based on Fused Hand Geometry
and Vascular Patterns, Sensors, 13 (3) (2013), 2895–2910.
https://doi.org/:10.3390/s130302895
148 B. Luna-Benoso, J. C. Martínez-Perales and J. Cortés-Galicia
[34] Ferrer, M. A., Morales, A., Ortega, L., Infrared hand dorsum images for
identification, Electronics Letters, 45 (6) (2009).
[35] Chen, X., Huang, M., Fu, Y., Simultaneous acquisition of near infrared image of
hand vein and pulse for liveness dorsal hand vein identification, Infrared Physics and
Technology, 115 (2021), 103688. https://doi.org/:10.1016/j.infrared.2021/103688
[36] Khan, M. H.-M., & Khan, N. M., Analysing Factors Affecting Hand Biometrics
during Image Capture, Procedia Computer Science, 32 (2014), 521–528.
https://doi.org/:10.1016/j.procs.2014.05.456
Received: August 25, 2021; Published: September 16, 2021