Indonesian Journal of Electrical Engineering and Computer Science

Vol. 29, No. 2, February 2023, pp. 1039~1046

ISSN: 2502-4752, DOI: 10.11591/ijeecs.v29.i2.pp1039-1064  1039

Design and implementation of low-cost vein-viewer detection

using near infrared imaging

Abdulrafa Hussain Maray1, Saba Qasim Hasan1, Naqaa Luqman Mohammed2

1
Department of Computer Systems Technologies, Mosul Technical Institute, Northern Technical University, Mosul, Iraq
2
Department of Computer Eng. Technical College, Northern Technical University, Mosul, Iraq

Article Info ABSTRACT

Article history: There are some medicines and medical treatments that need to be injected into
the human body through the blood vessels, and this requires placing the
Received Jul 18, 2022 cannula in the patient’s body. The blood vessels in the human body differ from
Revised Oct 13, 2022 one person to another, and medical personnel face major problems in finding
Accepted Oct 21, 2022 the blood vessels in most cases, because of The difference in skin color, where
it is difficult to see the blood veins in the skin with black pigment, and it is
difficult to find it in people with obesity because of the layers of fat, and in
Keywords: children and newborns because these veins are small. This study talks about
finding a way to photograph these veins, see them by design and
Canny implementation low cost prototype used equipment that were recycled old
Infrared cam device, such as a web camera, infrared lamps and overhead device, All of
Laplacian these devices are of low cost. Then process these images using binary image,
Prewitt histogram equalization, segmentation and threshold to detect these blood
Roberts veins. The algorithms for edges images detecting are many and complex, this
Vein viewer study used five methods to detect vein image, such as Sobel, Laplacian,
Canny, Roberts, and Prewitt.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Abdulrafa Hussain Maray
Department of Computer Systems Technologies, Mosul Technical Institute, Northern Technical University
AlMinsaa St, Mosul City, Nineveh Governorate, Iraq
Email: [email protected]

1. INTRODUCTION
Previous studies show that nearly more than ninety percent of patients in different countries of the
world requires peripheral cannulation of the intravenous route to perform laboratory tests in order to diagnose
diseases. Blood samples are taken from millions of people monthly. The economic vein viewer has
approximately four thousand five hundred dollars for small devices. While for the large devices, their material
cost reaches twenty-seven thousand dollars [1].
There are many advanced medical devices and equipment that can photograph veins, but they are not
available in hospitals because of their high price. In this study, a locally made device was design and
implementation low cost prototype device used equipment that were recycled. This device to detect blood veins
was manufactured in a simple and inexpensive way. One of the reasons that this study focuses on is solving
the problem of seeing the blood vessels in the human body, when the cannula is placed in the vein in order to
give patients gonorrhea, medical solutions, and needle glaucoma. People who have dark skin complexion,
it is difficult to identify and see the blood veins, as well as those who have many layers of fat, children and
newborns [2]-[10].
Table 1 shows a comparison between previous studies from (2010 to 2017). The average analytical
error rate is 61.5%. The researchers have been capable of lessen the mistake costs mentioned on this table in

Journal homepage: http://ijeecs.iaescore.com

1040  ISSN: 2502-4752

2017. The researchers had been capable of table for the proportion of 46.2%, while the percentage in 2010
grow to be approximately 77.1% [11]-[16].

Table 1. percentage of analytical errors

Author Year Percentage analytical error (%)
[11] 2017 up to 68.2
[12] 2017 up to 70.0
[13] 2015 60.0 to 70.0
[14] 2013 46.0 to 68.2
[15] 2012 46.0 to 68.2
[16] 2011 60.0 to 70.0
[17] 2010 77.10

2. THE PROPOSED METHOD

The following block diagram represents the most important material parts that were used in this model
show Figure 1. The first block in this diagram represents data acquisition such as infrared source, near-infrared
(NIR)-camera. The second block in this diagram represents processing methods such as edge detection,
segmentation, and classification. The final block in this diagram is the result for the vein finder.

Figure 1. Block diagram of the system design

2.1. Basic hardware composition design

The equipment has been recycled of the old overhead device, which was used in the presentation of
the presentation slides. Make use of lenses, body and moving levers. The distance is variable between the
camera and the infrared source is 13-37 cm, see Figure 2 design of the prototype device.
The first step of this diagram is to image acquisition from the infrared (IR)-camera, which needs an
infrared lamp at 100 Watt. This type of lamp is used in medical devices to treat the skin, as it gives a dim heat
that is used in the treatment and beautification of the skin. Figure 3 shows the best results between changing
the distance of the camera location and taking the most focus of the histogram IR-images. The intensity of this
IR-lamp is controlled by connecting a simple electronic circuit. This circuit consisting of a thyristor by pulse
width modulation (PWM) technology. Figure 4 shows the electronic circuit that shows how to control the
infrared lamp.

Figure 2. Prototype design

Indonesian J Elec Eng & Comp Sci, Vol. 29, No. 2, February 2023: 1039-1046
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1041

Figure 3. Distance of the IR-camera location and IR-images focus

Figure 4. Control circuit diagram for IR-source

The circuit is based on the principle of changing the root mean square (RMS) value of the voltage by
cutting off parts of the wave. The since wave with its positive and negative parts is (360) degrees, so only the
positive or negative part consists of (180) degrees, and if I say that I am firing at an angle of 90, then this means
that I am firing in the positive half of the wave and also in the negative half of the wave and so, as we increase
the angle of the IR-lamp to the triac (Q1), we lower the RMS voltage applied to the load, thus changing the
brightness of the IR-source. Where the RMS voltage is given by the (1) [17], [18]:
𝑉𝑚2
V 2 L(RMS)= [ 𝑖 2𝜋 − 2𝛼i + sin 2 𝛼] (1)
4𝜋

and we control the time required to open the diac (D5) (which will trigger the triac (Q1)) by changing the time
required to charge the capacitor, so changing the resistance will cause a change in the angle of the trigger
(alpha), which will change the RMS voltage leaving the circuit, see the Figure 5.
A webcam with a resolution of 15FPS images per second and a USB connection was used to connect
it to the computer. The filter inside it that blocks infrared rays from entering the camera was removed and
replaced with a filter to prevent the transmission of visible rays and allows the passage of infrared rays,
equipment has been recycled which is a small piece of old floppy disc, see the Figure 6.

Figure 5. Input and output signals (trigger at 90 degrees)

Design and implementation of low-cost vein-viewer detection using near … (Abdulrafa Hussain Maray)
1042  ISSN: 2502-4752

Figure 6. Internal component of web cam

2.2. Software methods

The tactics on this paintings consist of three main block diagram: first block is preprocessing, second
block is edge detection, and last block is segmentation. Inside the first block it contains of image acquisition,
histogram analyzing, and convert image data to gray method. Secondly block: the opposite technique of edge
detection changed into applied to part stumble on the vein pictures. The algorithms for edges images detecting
are many and complex, they used five methods to detect veins in the study such as: Sobel, Laplacian, Canny,
Roberts, and Prewitt. Inside the ultimate step, we followed the orientated bounding containers, the last block
represents the final results, the process of images segmentation and choosing a bounding for the final images.
see Figure 7.

Figure 7. Flowchart for evaluate system of vein edge detection

2.2.1. Preprocessing
After sampling the images with different levels, different lights, different ages and different skin
colors, treatments were carried out for these images and by designing algorithms to detect veins and arteries.
The (2) shows that the lowest value is zero and the highest value is 255 [18].
𝐼 − 𝐼 𝑚𝑖𝑛
J = 255. 𝐼 𝑚𝑎𝑥 − 𝐼𝑚𝑖𝑛 (2)

The infrared image is converted to grayscale to focus on important details and subtract the background of the
image. The threshold is selected after the spectrogram is performed by using histogram.

2.2.2. Edge detection method

In this study, many methods were used to determine the edges of the images, and these methods were
compared and the best ones were selected. Where the method Laplacian, Canny, Sobel, Roberts, and Prewitt
was chosen and these images are analyzed in these different ways to find the best results. All of these methods
are used on the gray image data, whose dimensions are (250×250).

2.2.3. Laplacian differential method

Laplacian operator is an isotropic operator, it is greater appropriate in satisfactory care approximately
the brink Feature without considering its surrounding pixel distinction. Laplacian operator responds to Isolate
pixel extra immoderate than to the threshold or line, so it's miles handiest carried out to the pics without noise.
In existed noise condition, it's far wanted low-bypass clear out before the use of above operator [7]. Laplacian
operator is defined as:

Indonesian J Elec Eng & Comp Sci, Vol. 29, No. 2, February 2023: 1039-1046
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1043

𝜕 2 𝑓 (𝑥,𝑦) 𝜕 2 𝑓 (𝑥,𝑦)
∇2𝑓(𝑥, 𝑦) = + (3)
𝜕 𝑥2 𝜕 𝑦2

∇2𝑓(𝑥, 𝑦) = 𝑓(𝑥 + 1, 𝑦) + 𝑓(𝑥 − 1, 𝑦) + 𝑓(𝑥, 𝑦 + 1) + 𝑓(𝑥, 𝑦 − 1) − 4𝑓(𝑥, 𝑦) (4)

2.2.4. Canny method

The Canny filter is one of the best and most widely used filters for defining the edges of images, as it
gives great results. The images taken from the camera have noise that negatively affects the edges of the images,
this noise must be eliminated and the image smoothed by using a Gaussian filter with the image. This step will
reduce the volume [18]-[21]. Gaussian function is:
1 ( x2 + y2 )
G(x, y) = exp (5)
2π α2 2 α2

𝜕𝐺
𝐺𝑛 = (6)
𝜕𝑛

𝑛 = [ 𝑐𝑜𝑠 ∅ 𝑠𝑖𝑛 ]𝑇 , 𝛻𝐺 = [ 𝜕𝐺⁄𝜕𝑋 𝜕𝐺⁄𝜕𝑋] (7)

when obtains the maximum = f (x, y) *Gn.

2.2.5. Sobel method

The Sobel operator is used to process the edges of images. The brightness intensity of each image is
calculated along with the contrast intensity of the image using a small, detachable filter with a correct value in
the horizontal and vertical directions. In concept the operator includes a pair of 3×3 convolution kernel as
shown under [18].
−1 0 1 1 0 1
Gx = −2 0 0 Gy = 0 0 0 (8)
−1 0 −1 −1 −2 −1

This design reaches the edges of the images more successfully, and the pixels are taken in the form of horizontal
and vertical, relative to the pixel grid [19]-[26].

𝐺 = √𝐺 2 𝑥 + 𝐺 2 𝑦 = 𝐺𝑥 + 𝐺 (9)

The direction of the edge angle which gives results upward is taken by (10).
𝐺𝑦
∅ = arctan ( ) (10)
𝐺𝑦

3. RESULTS AND DISCUSSION

In this study a system was designed and implemented for a system to detect veins from recycling old
devices, such as a web camera, an infrared lamp, an old overhead device, a floppy disk, and an electronic circuit
to control the intensity of the lighting. All these features made the design of the platform low cost.
Most of the previous studies use an IR-LED whose power does not exceed (10 Watt), and this low
power cannot penetrate the skin to a large thickness, so the infrared source is placed from the same side of the
camera. They are linked in one site. Their results will be limited. As for this innovative design increased power
of the infrared source to (100 Watt), and intensity of the lighting is controlled by an electronic circuit to get the
best results see Figure 8.
The best result obtained is at a distance of 16 cm from the camera and 90% of the energy of the IR-
lamp. The Figure 9 shows the image of the infrared camera with histogram for chosen the threshold filter point.
In this analysis points are threshold chosen between 150-66 to get the best results.
The Figure 10 shows the difference between the techniques used in detecting the angle in images. The
algorithms for edges images detecting are many and complex. They used five methods to detect veins such as
Figure 10(a) Sobel, Figure 10(b) Laplacian, Figure 10(c) Canny, Figure 10(d) Roberts, and Figure 10(e)
prewitt. The Figure11 shows classification result, these (1 to 7) circles represent the best places to use the
cannula. This image represents the best result obtained. It is possible to specify more points by changing the
filter values, or by changing the illumination values of the infrared source. All of these points that were detected
in this study are valid for using the medical cannula or taking blood samples from the patient's body.

Design and implementation of low-cost vein-viewer detection using near … (Abdulrafa Hussain Maray)
1044  ISSN: 2502-4752

Figure 8. Beast results of IR-image focusing

Figure 9. Infrared image acquisition and histogram

(a) (b) (c) (d) (e)

Figure 10. IR-image with edge detection result; (a) Sobel, (b) Laplacian, (c) Canny,
(d) Roberts, and (e) Prewitt

Figure 11. Edge detection and classification

4. CONCLUSTION
The economic vein-viewer has an estimated value of approximately $4,500. For the mobile type, about
$27,000 for the non-mobile type. In this study, a system was designed, priced at only $25. This price is
considered very low for the efficiency it performs. Primarily based on the review, the low-cost vein-viewer
development system designee using near infrared imaging, in the segment of improvement, in particular,

Indonesian J Elec Eng & Comp Sci, Vol. 29, No. 2, February 2023: 1039-1046
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1045

targeting its goal to reduce the instances of ignored peripheral subcutaneous veins in the course of blood
collection and intravenous insertion for medicine. Common parameters used in evaluating the device were the
following, color skin, venipuncture, temperature and age, the illumination of the ambient place and the intensity
of the infrared source. When using a set of edge detection filters, the best results were obtained in a filter Sobel
and Roberts. The edge detection performance for the Canny 55.6%, Laplacian 42.89%, Robert’s 95.51%,
prewitt 87.41%, and Sobel 97.21%.

REFERENCES
[1] G. Cantor-Peled, M. Halak, and Z. Ovadia-Blechman, “Peripheral vein locating techniques,” Imaging in Medicine, vol. 8, no. 3, pp.
83-88, 2016.
[2] J. Zhang and J. Yang, “Finger-vein image enhancement based on combination of gray-level grouping and circular gabor filter,”
2009 International Conference on Information Engineering and Computer Science, 2009, pp. 1-4, doi:
10.1109/ICIECS.2009.5363431.
[3] A. H. Maray, O. I. AlSaif, and K. H. Tanoon “Design and implementation of low cost medical auditory system of distortion using
Microcontroller,” Journal of Engineering Science and Technology. vol. 17, no. 2, pp. 1068-1077, 2022.
[4] B. S. Mohamad-Ali, and A. H. Maray “Low coast medical data acquisition system using AVR-MCU and PC,” Journal of
Theoretical and Applied Information Technology, vol. 42, no. 2, pp. 217-220, 2012.
[5] C.-B. Yu, D.-M. Zhang, H. -B. Li and F. -F. Zhang, “Finger-vein image enhancement based on muti-threshold fuzzy algorithm,”
2009 2nd International Congress on Image and Signal Processing, 2009, pp. 1-3, doi: 10.1109/CISP.2009.5303807.
[6] J. Yang and J. Wang, “Finger-vein image restoration considering skin layer structure,” 2011 International Conference on Hand-
Based Biometrics, 2011, pp. 1-5, doi: 10.1109/ICHB.2011.6094320.
[7] W. Pi, J. Shin and D. Park, “An effective quality improvement approach for low quality finger vein image,” 2010 International
Conference on Electronics and Information Engineering, 2010, pp. 424- 427, doi: 10.1109/ICEIE.2010.5559667.
[8] M. G. Saeed, K. H. A. Faraj, M. T. Alasaady, and F. L. Malallah. “Three levels quality analysis tool for object oriented
programming,” International Journal of Advanced Computer Science and Applications, vol. 9 , no. 11, pp. 522-536, 2018.
[9] M. Es-Sabry, N. El Akkad, M. Merras, A. Saaidi, and K. Satori, “A new color image encryption algorithm using multiple chaotic
maps with the intersecting planes method,” Scientific African, vol. 162022, doi: 10.1016/j.sciaf.2022.e01217.
[10] K. I. Ahmed, M. H. Habaebi, and M. R. Islam, “A real time vein detection system,” Indonesian Journal of Electrical Engineering
and Computer Science(IJEECS), vol. 10, no. 1, pp. 129–137, 2018, doi: 10.11591/ijeecs.v10.i1.pp129-137.
[11] J. West, J. Atherton, S. J. Costelloe, G. Pourmahram, A. Stretton, and M. Cornes, “Preanalytical errors in medical laboratories: a
review of the available methodologies of data collection and analysis,” Annals of Clinical Biochemistry: International Journal of
Laboratory Medicine, vol. 54, no. 1, pp. 14-19, 2017, doi: 10.1177/0004563216669384.
[12] D. Najat, “Prevalence of pre-analytical errors in clinical chemistry diagnostic labs in Sulaimani City of Iraqi Kurdistan,” PLOS
ONE, vol. 12, no. 1, 2017, doi: 10.1371/journal.pone.0170211.
[13] M. Salinas et al., “Ten years of preanalytical monitoring and control: synthetic balanced score card indicator,” Biochemia Medica,
pp. 49–56, 2015, doi: 10.11613/BM.2015.005.
[14] S. Patra, B. Mukherjee, and A. K Das. “Pre-analytical errors in the clinical laboratory and how to minimize them,” International
Journal of Bioassays, vol. 2, no. 3, pp. 551–553, 2013.
[15] J. A. Hammerling, “A review of medical errors in laboratory diagnostics and where we are today: table 1,” Laboratory Medicine,
vol. 43, no. 2, pp. 41-44, 2012, doi: 10.1309/LM6ER9WJR1IHQAUY.
[16] G. Lippi et al., “Preanalytical quality improvement: from dream to reality,” Clinical Chemistry and Laboratory Medicine, vol. 49,
no. 7, 2011, doi: 10.1515/CCLM.2011.600.
[17] B. Goswami, B. Singh, R. Chawla, and V. Mallika, “Evaluation of errors in a clinical laboratory: a one-year experience,” Clinical
Chemistry and Laboratory Medicine, vol. 48, no. 1, 2010, doi: 10.1515/CCLM.2010.006.
[18] P. Wayalun, N. Laopracha, P. Songrum, And P. Wanchanthuek, “Quality evaluation for edge detection of chromosome gband
images for segmentation,” Applied Medical Informatics, vol. 32, no. 1, pp. 25-32, 2013.
[19] M. T. Alasaady, M. G. Saeed and K. H. Faraj, “Evaluation and comparison framework for data modeling languages,” 2019 2nd
International Conference on Electrical, Communication, Computer, Power and Control Engineering (ICECCPCE), 2019, pp. 68-
73, doi: 10.1109/ICECCPCE46549.2019.203750.
[20] R. R. O. Al-Nima, S. Q. Hasan, and S. Esmail, “Exploiting the deep learning with fingerphotos to recognize people,” International
Journal of Advanced Science and Technology, vol. 29, no. 7, pp. 13035–13046, 2020.
[21] K. H. Thanoon, S. Q. Hasan, and O. I. Alsaif, “Biometric information based on distribution of arabic letters according to their
outlet,” International Journal of Computing and Digital Systems, vol. 9, no. 5, pp. 981-991, 2020, doi: 10.12785/ijcds/090518.
[22] A. P. Moholkar, D. S. Bankar, and P. B. Karandikar, “Synthesized carbon material with semi-organic and aqueous electrolyte for
supercapacitor application,” Indonesian Journal of Electrical Engineering and Computer Science (IJEECS), vol. 28, no. 2, pp. 654-
662, Nov. 2022, doi: 10.11591/ijeecs.v28.i2.pp654-662.
[23] R. Verayiah and A. Mohamed, “Comparison of weak load bus detection using LQP_LT index with PV and QV analysis of PSS/E,”
Indonesian Journal of Electrical and Computer Science (IJEECS), vol. 12, no. 2, pp. 577-584, 2018, doi:
10.11591/ijeecs.v12.i2.pp577-584.
[24] M. S. A. Mustaza, M. A. M. Ariff, and S. N. Ramli, “An extensive review of energy storage system for the residential renewable
energy system,” Indonesian Journal of Electrical Engineering and Computer Science (IJEECS), vol. 18, no. 1, pp. 242-250, 2020,
doi: 10.11591/ijeecs.v18.i1.pp242-250.
[25] S. Q. Alhashmi, K. H. Thanoon and O. I. Alsaif, “A proposed face recognition based on hybrid algorithm for features extraction,”
2020 6th International Engineering Conference “Sustainable Technology and Development" (IEC), 2020, pp. 232-236, doi:
10.1109/IEC49899.2020.9122911.

Design and implementation of low-cost vein-viewer detection using near … (Abdulrafa Hussain Maray)
1046  ISSN: 2502-4752

[26] N. L. Mohammed, M. S. Aziz, and O. I. AlSaif, “Design and implementation of robot control system for multistory buildings,”
TELKOMNIKA (Telecommunication Computing Electronics and Control), vol. 18, no. 5, pp. 2682–2689, 2020, doi:
10.12928/telkomnika.v18i5.15338.

BIOGRAPHIES OF AUTHORS

Abdulrafa Hussain Maray he received his B.Sc in computer engineering from

Northern Technical university, Mosul, Iraq at 2003 and the M.Sc in Microelectronic from
computer Technical engineering department Northern Technical university at 2013. He is
currently working as lecturer in Mosul Technical Institute/Northern Technical University in
Mosul Iraq. He can be contacted at email: [email protected].

Saba Qasim Hasan received her B.Sc in computer science department Mosul
University, Mosul, Iraq at 1996 and the M.Sc in Image Security Processing from Computer
Science department Mosul university at 2003. She is currently working as lecturer in Mosul
Technical Institute Northern Technical University in Mosul Iraq. Her research interests image
security and objects detection. She can be contacted at email: [email protected].

Naqaa Luqman Mohammed received his B.Sc in Computer Engineering from

Northern Technical university, Mosul, Iraq at 2003 and the M.Sc in Image Security
Processing from Computer Engineering from Northern Technical university in 2011. She is
currently working as lecturer in Computer Engineering Northern Technical University in
Mosul Iraq. Her research interests’ control, arduino, and FPGA. She can be contacted at
email: [email protected].

Indonesian J Elec Eng & Comp Sci, Vol. 29, No. 2, February 2023: 1039-1046