Journal of Physics: Conference

Series

PAPER • OPEN ACCESS You may also like

- A mechanical arm based on image
Method of peripheral vein detection using recognition and remote control
Kangmin Li
Electrical Impedance Tomography - CATALOG OF OBSERVED TANGENTS
TO THE SPIRAL ARMS IN THE MILKY
WAY GALAXY
To cite this article: M A San-Pablo-Ju´arez and D M Orozco-Corona 2021 J. Phys.: Conf. Ser. 2008 Jacques P. Vallée
012017
- Home-Based Online Multisensory Arm
Rehabilitation Monitoring System
Zakarya Farea Shaaf, Muhammad Mahadi
Abdul Jamil, Nayef.A.M. Alduais et al.

View the article online for updates and enhancements.

This content was downloaded from IP address 201.166.191.49 on 20/07/2024 at 01:58

4th Latin American Conference on Bioimpedance 2021 (CLABIO 2021) IOP Publishing
Journal of Physics: Conference Series 2008 (2021) 012017 doi:10.1088/1742-6596/2008/1/012017

Method of peripheral vein detection using Electrical

Impedance Tomography
M A San-Pablo-Juárez1 and D M Orozco-Corona2
1
Escuela de Ingenierı́as, Universidad de las Américas Puebla. Ex Hacienda Sta. Catarina
Mártir S/N. San Andrés Cholula, 72810 Puebla. México.
2
División de Tecnologı́as de la Automatización e Información, Universidad Tecnológica de
Querétaro. Av. Pie de la Cuesta 2501, Nacional, 76148 Santiago de Querétaro, Qro. México.
E-mail: [email protected]

Abstract. The aim of the current contribution was to develop a method for locating arm veins
in patients by using an electrical impedance tomography. Eight surface electrodes were placed
on a simulated arm, separated at a fixed and equal distance from each other. The electrodes fed
information about electrical impedance (conductivity) to EIDORS software. The three main
phases (veins, muscle, and bone) each have different conductivity, which the device and software
convert into a colour-coded tomograph displayed on a screen. The colour for blood on the image
is previously set (in this case, red), revealing the location of the veins in the selected section
of the arm. A sample run was made, successfully finding the correct location and size of the
modelled vein, thus demonstrating the accurate operation of the device.

1. Introduction
Of the approximately 500 million venipunctures carried out annually in the world, 2.7-4.8% are
not performed successfully on the first attempt, according to recent studies. This translates into
around 15,000 patients per day that are subjected to unsuccessful perforations for the extraction
of blood or insertion of fluids, causing unnecessary discomfort and pain [1, 2].
Generally, the method of palpation and visual identification is used to find a vein to be
punctured. However, some factors make venipuncture difficult, including overweight, vascular
tone, vein diameter, and varicose veins. The skill of the person puncturing the vein may also be
relevant [1].
Some devices have been designed to solve the problem of locating peripheral veins of the
arm, forearm and hand. For instance, infrared detectors [3, 4, 5]show veins with near infrared
light (700-1000 nm) [6, 7]. Additionally, high-resolution ultrasound scanners provide a very
good quality image in real time. Although these instruments are effective, they are relatively
expensive. The venography technique can also produce images, but requires the injection of a
contrast agent.
A novel and economical technique is electrical impedance [1], which was presently tested with
surface electrodes to locate veins by means of the 4-point method with the injection of a fixed
electrical current.
In this work, a simulation of an alternative method to locate peripheral veins that uses
Electrical Impedance Tomography (EIT) was made, by means of 8 Ag/AgCl surface electrodes.
Unlike the simple impedance methods [1, 8], the proposed method displays an EIT image of the

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
4th Latin American Conference on Bioimpedance 2021 (CLABIO 2021) IOP Publishing
Journal of Physics: Conference Series 2008 (2021) 012017 doi:10.1088/1742-6596/2008/1/012017

Figure 1. Electrical Impedance Tomography technique.

measured section with high contrast on a square screen. Conductivity data from the electrodes
was fed into EIDORS software previously set with reported values of vein, muscle, fat, skin,
and bone impedance (conductivity). Electrical impedance tomography (EIT), employed in civil
engineering to map terrain on a small scale, is based on a tomographic image reconstructed with
EIDORS software [9, 10, 11].

2. Materials and methods

2.1. Electrical Impedance Tomography
EIT is an economical non-invasive imaging technique that can be a valuable tool for medical
diagnosis [8]. For image construction with EIT, a low frequency electrical current is applied
to an object and the resulting electric field on its surface is measured (Figure 1). The pattern
produced by the injected electrical current allows a series of electrodes to make a set of readings
of electrical potential, which are recorded from the boundary of the conductive region [10]. This
data on the distribution of conductivity is fed into a computer or EIDORS software to generate
an image [9]. Although a computer furnishes a very accurate topograph from the measurements
made of the overall distribution of conductivity [11, 12, 13, 14], the most common way of
displaying images is with EIDORS software [15, 16, 17].
The main equation for EIT is [8, 15]:

∇(σ∇φ) = 0 ∈ D (1)
Where ∇ is the gradient operator, ∇φ represents the static electric field, σ is the conductivity
of the body, φ expresses the electric potential, and D is the body to be imaged [8, 15]. There
are two processes involved in EIT, a forward problem and an inverse problem. In the forward
problem, the initial conditions of conductivity are proposed, a current I is injected to find a
potential distribution, and equation (1) is resolved to calculate the potential φ(x, y) ∈ D. In the
inverse problem, an error between the computed potential φ(x, y) and the experimental value in
the forward problem is compared until it is minimized [15, 16].
For the EIT technique, AC electrical current is injected into each pair of electrodes in region
D in order to measure a set of potentials between them (1). The distribution of conductivity
inside of region D is given by equation (2):

σ = σreal + i σcomplex (2)

where σreal corresponds to the conductivity of the material and the σcomplex corresponds to
its permittivity. In this study, only the σreal of equation (2) was reconstructed because of the
nature of the phases inside the arm. While veins and muscle are very conductive, bone is not.
The permittivity of the phases was not used.

2
4th Latin American Conference on Bioimpedance 2021 (CLABIO 2021) IOP Publishing
Journal of Physics: Conference Series 2008 (2021) 012017 doi:10.1088/1742-6596/2008/1/012017

Figure 2. EIT with an open electrode configuration.

Figure 3. Electrode position and display of the location of veins near the surface of the arm.

The typical frequencies involved in EIT are above 1 M Hz, with a peak current above 0.5 mA
[15, 18].

2.2. Linear EIT with 8 electrodes

Most applications of EIT are for a circular distribution or region (see Figure 2). Accordingly, a
ring of electrodes is attached around the surface area to be analyzed. However, the measurement
of the current study was made at the surface of the skin, making open linear distribution more
appropriate. The EIT obtained from this configuration is only a section of that resulting from
the electrode ring configuration of a circular distribution. Since the present region of interest
(a part of the arm) for the tomographic image is relatively small in size (no more than a 10 cm
long), the proposed number of electrodes is only 8. A larger number would not be feasible.
Due to the square configuration of the screen a finite element grid of 360 elements was
considered to solve the direct and inverse problem in the modeling. So a square configuration is
used to model and solve the EIT in this work. Adjacent pattern stimulus for injection of current
I, adjacent measurement of the potential φ and 0.5 mA peak (1 mApp) were used.
A simplified model for the linear reconstruction method is expressed in equation (3):

[σ]360×1 = [J]360×40 [φ]40×1 (3)

Where σ is the vector of conductivity for the inverse problem of the image reconstruction
method, J is the Jacobian matrix of the system, and φ is the vector of the measured potentials
of the surface [14]. In this case, σ has size 360 × 1 because the finite element mesh on EIDORS
software contains 360 triangular elements; φ is the measurement potentials vector, that contains

3
4th Latin American Conference on Bioimpedance 2021 (CLABIO 2021) IOP Publishing
Journal of Physics: Conference Series 2008 (2021) 012017 doi:10.1088/1742-6596/2008/1/012017

40 independent measurements because a maximum of 5 independent measurements are made

in an 8 electrode configuration for an adjacent current injection [15], it makes a total of 5 × 8
independent measurements; and finally the jacobian (J) size is 360× 40 due to the mathematical
model configuration.
A practical example of the scanner position for sensing the potentials involved in EIT imaging
is illustrated in Figure 3. The image of the veins is displayed on the screen of the EIT scanner.

2.3. EIDORS
EIDORS (electrical impedance tomography and diffuse optical tomography reconstruction
software) is designed to generate EIT images [17]. It processes a finite number of potential
measurements of the conductivity and permittivity of an object and displays an image of
electrical impedance distribution on a screen. The conductivity distribution was employed
in the current study because blood conductivity (0.67 S/m) contrasts with that of muscle
(0.23 − 0.63S/m), fat (0.036S/m), bone (0.006S/m) and skin (0.00012S/m) [19]. The 3.10
with Netgen 5.3 version of EIDORS was presently run on MATLAB c . It contains a lot of
demos in m language.

2.4. Method for scanning and displaying the position of the veins
After considering all the aspects of EIT methodology, the most appropriate ones were selected
to design a new scanner for locating surface veins. The procedures are described hereafter.

(i) Eight Ag/AgCl electrodes are attached to the surface of the skin to detect the location of
the veins in that area. It is very important to ensure contact of the electrode with the
skin for the measurement of electrical potential. The distance k (in cm) between electrodes
(e.g., 0.5cm) and the radius r of the electrodes must be taken into account.
(ii) The position of the electrodes is established to normalize the scanned region to a rectangular
model of a cm × b cm (example: 4cm × 3cm).
(iii) The EIT technique for the detection of conductivity is applied (real part measurements of
impedance), and the image is reconstructed on EIDORS software. The parameters proposed
for this method are: current = 1 mApp, frequency = 1 KHz (or some typical value used in
EIT), with a sinusoidal pattern. The device is held at 0◦ and used with the synchronous
demodulation method.
(iv) The EIT of the evaluated region is displayed as a colour-coded map on a screen, with each
colour portraying a certain magnitude of conductivity.
(v) The desired vein is located on the screen in the colour representing the conductivity of
blood.
(vi) The electrode position is the reference point for locating the vein.

EIDORS software reconstructs the tomography based on the surface measurements of

electrical potential with the mathematical model chosen and the values established for the
injection of current. Other configuration parameters are at the default settings, considering
the finite element method used to solve the forward problem, the triangular shape of the finite
elements, iteration procedures, and linear and nonlinear reconstruction methods. The EIT
technique requires a physical EIT tomograph.

3. Results
The implementation of the method herein designed was simulated with EIDORS software run
on MATLAB c . The following parameters were used: an electrode contact impedance of
0.01 S/m, 8 electrodes, a 2 − mm separation of the electrodes, and an adjacent injection of

4
4th Latin American Conference on Bioimpedance 2021 (CLABIO 2021) IOP Publishing
Journal of Physics: Conference Series 2008 (2021) 012017 doi:10.1088/1742-6596/2008/1/012017

Figure 4. Vein detection with EIT.

current. Additionally, conductivity was normalized for muscle (1S/m), blood (1.44S/m), and
veins (0.776S/m) and the device was prepared with the following settings: bones (modelled
= 2), muscle (modelled = homogeneous), veins (modelled = 1).
The forward problem was modelled with a finite element mesh, two circular areas for bones,
and a vein inside a muscle region, as found in a human arm. The inverse problem was solved
with the Jacobian method on EIDORS to achieve an EIT where the vein is clearly visible and
easily located with high contrast. The EIT display clearly showed the position of the vein, as
illustrated in Figure 4.
A map of the three main phases of conductivity (generated with EIDORS software) was
viewed in high contrast and printed for reference (Figure 4). The position of the vein in
the reconstructed EIT image corresponds to the position of the modelled vein. The electrical
impedance of the vein is clearly distinguishable from that of other phases, such as muscle.
The area determined for the simulation presented in this work was 15 cm in the y and 16 cm
in the x axes for the forward solution, but only a mesh of 10 cm in the y and 12 cm in the x
axes was considered for the inverse problem. The cross-sectional area for the simulated vein
was 19.6 × 10−6 m2 , reflecting a modelled radius of 2.5 mm. Hence, the EIT imaging method
presents an accurate display for locating veins of this size.
The technique under study produces the image of the position of the vein even when the high
impedance of the skin is considered (see Figure 5). The position of the vein is clearly identifiable
due to its contrast with other tissues of the forearm. Even though the contrast increased for the
vein, it decreased for the bones. High-impedance noise also appears in the image.

4. Discussion
The results of the simulations show that it is possible to locate the position of veins with EIT.
As veins are low impedance bodies (high electrical conductivity), they are distinguishable from
bone, muscle, and fat, and therefore detectable on the display of EIT. The high contrast of
the tomographic image accurately shows the vein in its modelled position. One of the factors
justifying the development of this method of vein detection is that the venopunctions are difficult
in obese patients, therefore not only the bone and muscle but also fat is considered in the
simulation. When the skin is considered in the simulation, the reconstructed image shows that
the vein is still distinguishable in the image, but other high impedances introduce some noise to
the image. The vein remains recognizable due to its low impedance, which is easily located on

5
4th Latin American Conference on Bioimpedance 2021 (CLABIO 2021) IOP Publishing
Journal of Physics: Conference Series 2008 (2021) 012017 doi:10.1088/1742-6596/2008/1/012017

Figure 5. Vein position inside of the arm when skin impedance is considered for the EIT
technique.

the image due to its contrasting color on the impedance tomography. Since the objective was to
establish a method to visualize the position of veins on a square screen scanner, EIT is presented
in square geometry. However, the EIT technique can be implemented without much difficulty in
a circular or elliptical geometry as in generic applications of the technique. However, it should
be noted that the circular geometry type is widely reported in the literature as one of the main
geometries.

5. Conclusions
A new method was tested for the detection of peripheral veins based on EIT. The data
captured by the electrodes was converted into a tomograph with EIDORS software. The distinct
conductivity of the veins versus that of muscle, fat and bone resulted in a high-contrast image
displaying a vein of 19.6 × 10−6 m2 in the correct position. The next step is to evaluate the new
method with electrodes on a real arm to produce a real EIT tomograph.

References
[1] Al-Harosh M and Shchukin S 2017 J. Electr. Bioimp. 8 79–83
[2] Ganesh S 2007 Depth and size limits for the visibility of veins using the veinviewer imaging system. Master’s
thesis University of Tennessee and University of Memphis.
[3] Lee J, Moon S, Lim J, Kim K, Lee J H, Gwak M J and Kim K S 2016 A finger-vein imaging and liveness
detection for identity authentication using 2-axis mems scanner. 2016 International conference on Optical
Mems and Nano photonics (OMn).
[4] Jayaprabha T and Suresh A 2016 Blood flow, vein and nerves detector using an nir sensor with estimation
for embedded signal processing. 2016 International Conference on advanced Control and Computing
Technologies (ICACCCT).
[5] Al-Ghozali H, Setiawardana and Sigit R 2016 Vein detection system using infrered camera. 2016 International
electronics symposium (IES).
[6] Pan C T, Francisco M, Yen C K, Wang S Y and Shiue Y L 2021 Sensors 19 URL Doi:10.3390/s19163573
[7] Ahmed K, Habaeibi M and Islam M 2018 Indonesian journal of electrical Engineering and Computer Science.
10 URL DOI: 10.11591/ijeecs.v10.il
[8] Zhang X, Xu G, Zhang S, Li Y, Guo Y, Wang Y and Yan W 2014 IEEE Trans. on Magnetics.
[9] Molinari M 2002 High fidelity imaging in electrical impedance tomography. Ph.D. thesis University of
Southhampthon.
[10] Polydorides N 2002 Image reconstruction algorithms for soft-field tomography. Ph.D. thesis University of
Manchester Institute of Science and Technology.

6
4th Latin American Conference on Bioimpedance 2021 (CLABIO 2021) IOP Publishing
Journal of Physics: Conference Series 2008 (2021) 012017 doi:10.1088/1742-6596/2008/1/012017

[11] Adler A 1995 Measurement of pulmonary function with electrical impedance tomography Ph.D. thesis
Universit de Montral
[12] Li T, Isaacson D, Newell J and Saulnier G 2013 Journal of Physics Conference Series.
[13] Cheney M, Isaacson D and Newell J 1999 SIAM review. 41 85–101
[14] Polydorides N and Lionheart W 2002 Measurement Science and Technology
[15] Holder D 2005 Institute of Physics Publishing.
[16] Lionheart W 2005 Physiol. Meas.
[17] Adler A and Lionheart W 2006 Physiol Meas. 27 S25–42 URL doi: 10.1088/0967-3334/27/5/S03
[18] Olson W H 2021 Medical Instrumentation Application and Design: Electrical Safety (John Wiley and Sons)
chap 14, pp 638–674 4th ed
[19] Mueller J and s Siltanen 2012 Linear and nonlinear inverse problems with practical applications. (Society for
Industrial and Applied Mathematics.)