2013 International Conference on Signal Processing, Image Processing and Pattern Recognition [ICSIPRI

Real-time law cast infrared vein imaging system

1 2
Manam Mansoor , Sravani.S.N , Sumbul Zahra Naqvi, Imran Badshah , Mohammed Saleem
1 - manam_59 @yahoo.com ; 2- sreeramsravani @ gmai1.com
Department ofMedical Electronics
MS.Ramaiah Institute ofTechnology
Bangalore, Karnataka, India

Abstract-Vein detection is one of the latest biomedical A. Needfor vein detection

techniques researched today. While the concept behind the
method is simple, there are various challenges to be found
throughout the design and implementation of a device concerning
the lighting system and the image processing algorithms at a very
low price. While a very few devices based on the infrared
technique have been implemented, there still exists a strong need
to develop such medical devices. The major problem faced by the
doctors today is difficulty in accessing veins for intra-venous
drug delivery. With improper detection of
problems like bruises, rashes, blood clot etc. occur. Therefore a
non-invasive subcutaneous vein detection system
developed success-fully based on near infrared imaging and
interfaced to a laptop. A customized webcam is used for
capturing the vein images and Computer Vision is used for the
processing. The pilot study details are also provided in this
paper. This also has application in treatment of varicose veins,
deep vein thrombosis and vascular ailments.
characteristics of this device discussed in
portability and low cost.
Index Terms-vein imaging, infra-red imaging,
threshold, contrast stretching, low cost imaging
I. INTRODUCTION
A huge progress is made in the field of biomedical imaging
technologies but they have generally resulted in high product
costs. Another major concem is the portability factor. We have
addressed these two major concems in this device to a
maximum extent. As compared to infrared imaging, significant
amount of work is done in other areas of imaging techniques.
The use of infrared imaging technique is a relatively less
explored area but promises to deliver high-end results at low
development costs. The major clinical problem faced by the
physicians is difficulty in accessing veins for intra-venous drug
delivery and other purposes. It is also noted that improper
location of veins is a cause for delay in several medical
treatments. It is relevant in case of paediatrics, obese, dark
toned people and also in adult patients. Unnecessary
puncturing of veins is done by the physicians because the
visibility of veins is not clear. Therefore this causes various
problems to the patients and especially in children and aged.
The results are swelling, irritation, bleeding and blackening of
the skin. Although a significant amount of work has been done
in this area and devices like Accuvien VeinVeiwer have
come up, but the major problem lies in their cost and
portability factor.
When the doctors are treating trauma patients, every
second counts. Bruises, bums and other physical injuries make
it difficult to locate veins and administer life saving drugs or
solutions. In such cases it becomes very necessary to have a
device that detects the exact location of required vein. Also in
case of blood transfusion, blood donation, blood withdrawal,
etc. it is necessary to know the exact position of the veins.
veins, several Even trained nurses and doctors many times fmd it difficult to
exactly locate the blood veins, on the first attempt, especially
has been for obese people. In various medical situations, the exact
location of veins needs to be identified. The other situations
where vein imaging is required are:
Intravenous injections: For giving medicines and drugs to
the patients, intra- venous injections are given by doctors
The two key and nurses.
the paper are Bruises and Bums: In case of vein diseases like Deep Vein
Thrombosis and Varicose Veins, bruises appear on the
skin, therefore for the treatment of these diseases,
adaptive
detection of veins is highly essential. Accidents involving
first or second degree of bums cause the scarring of the
skin. Here appearance of the skin becomes deterred
causing the skin to appear whiter or in certain cases darker.
The determinations of veins become tough in such cases as
weil.
Blood transfusions: It is a process in which blood is given
to the person intravenously. Blood donation, kidney
dialysis also need perfeet vein detection
Among children: Locating veins in young children and
infants may be especially difficult and having to puncture
them several times with a needle is very frightful and
agonizing for the child.[4]
Geriatrics: Many elderly people often require numerous
blood tests or medicinal injections and an efficient means
of puncture would reduce excessive bruise and enhance
the patients overall comfort level.
The wrong administration of intra-venous injections will lead
to several problems like swelling of the skin, appearance of
rashes ,damage to bones and nerves, c10tting of blood allergie
reactions ,appearance of black spots , darkening of veins due to
peripheral venous scarring ,arterial damage which may result in
haemorrhage and gangrene.

978-1-4673-4862-1/13/$31.00 2013 IEEE

 2013 International Conference on Signal Processing, Image Processing and Pattern Recognition [ICSIPRI 2

11. NEAR INFRARED IMAGING

Human eyes can only detect visible light that occupies a
very narrow band (400 - 700nm) of the entire electromagnetic
spectrum [2].However, there is much more information
contained in other bands of the electromagnetic spectrum
rejected by the objects of interest. For human vein patterns on
the periphery, the visibility under normal visible light
conditions is very low. However, this can be resolved by using
Near-Infrared imaging techniques.

The special attributes of Near-Infrared imaging which

makes it suitable for vein detection are:
NIR can penetrate into the biological tissue up to 3mm of
depth, [3]. Figure 1: LED Ring with driver

The reduced haemoglobin in venous blood absorbs more

of this infrared radiation than the surrounding tissue, [2].
Therefore, by exposing desired body part with the infrared
radiation of specific wavelength, the vein image can be
captured by an IR camera(camera used in our system is as
shown in Figure2). In the resulting image, the veins appear
darker than the surrounding tissue, biologically; there is a
medieal spectral window which extends approximately from
about 700 to 900 nm, where light in this spectral window
penetrates deeply into tissues, thus allowing non-invasive
investigation. The infrared radiation is absorbed in a different
way in various types of tissues. In order to achieve visual
penetration through the respective tissue, lighting should be
performed under a very tight optical window, namely 740 nm
up to 960 nm (inside the near infrared part of the Figure2 : Modified web camera
electromagnetic radiation spectrum).Hence in this device near
infrared technique is chosen rightly to view the veins.
IV. ALGORITHM

111. DEVICE SETUP

Our acquisition system (Figurei) is composed of a
modified wed camera with a long-pass interferometric filter,
a NIR ring of LEDs is used for illuminating the desired body
part with infrared light. The LED ring has a circuit driver and
the power is provided from the laptop itself through a USB
cable [10] the camera (Figure2) is also interfaced through a I----?I DISPLAY IMAGE
USB cable. Our choice was based on the appropriate spectral
response of the CMOS sensor in the NIR range of light, its
compactness and the communication protocol allowing both DISPLAY IMAGE
video signal transfer and hardware configuration and also the
cost. Once the signals are sent to the laptop the processing is
t-----?I DISPLAY IMAGE
done and the results are displayed on the laptop screen. The
detailed list of the components is provided in tablel.
The software used for processing is Computer
Vision(OpenCV) on Ubuntu. It is an open source computer
vision library which is originally developed [1]. The library is The algorithm of the image processing is as shown above
written in C and C++ and runs under Linux, Windows and Mac and the details of each step is as follows (here the sampie of
OS X. The OpenCV library contains over 500 functions that one frame is shown but on the device it works on real-time)
span many areas in vision, including factory product
inspection, medical imaging, security, user interface, camera
calibration, stereo vision, and robotics. The use of this reduces
the cost immensely as it is available for free and gives efficient
platform for user.
 2013 International Conference on Signal Processing, Image Processing and Pattern Recognition [ICSIPRj 3

A. IMAGE ACQUISTION: adaptive thresholding is performed as it has several advantages

The vein imaging camera detects light that includes even over the normal histogram equalization.
the Infrared wavelengths crucial to discerning the veins. This
imaging is done in real-time, which implies that the output is
displayed as a video. And video as we know is nothing but a
sequence of still images called as "frames". Hence processing
this acquired coloured video to present the veins more
distinctly is carried out on individual frames.

Figure4 : After contrast stretching and adaptive thresholding

E.SEGMENTATION:
This stretched gray scale is now ready to be segmented.
Using one of the simplest of segmentation, the Adaptive
thresholding marks clear boundaries between "foreground" and
Figure3 : After noise removal and smoothing
"background" pixels determined by the threshold value opted
by the user through the track-bar. This segmentation hinders
B. GRA Y SCALE CONVERSION:
the speed of processing the least and has so far resulted as the
The first step of processing the image is to enhance the
most reliable. The constant parameter is introduced to
picture quality. And by this, we mean discerning the pixels, or
compensate for the drawbacks of the early adaptive
in layman's terms, enhance the contrast. Converting to gray
thresholding [8]. Upon this segmentation, the image results as a
scale allows for faster processing in the further stages as
binary image demarcating quite clearly the foreground image
compared to coloured images and still yield discernible
and the background image based on the contrast enhanced
visualization of the veins in real-time. Smoothing and Noise
image.[9]
removal.

C.SMOOTHING AND NO/SE REMOVAL:

The real-time data acquired by the camera contains noise.
Smoothing is generally done to remove noise so that the
underlying information can be revealed. In order to eliminate
noise, such as salt & pepper noise from the vein images a 3x3
Median filter has been used. The median filter is a non-linear
filter and replaces each pixel by the median of neighbouring
pixels.[12]
It is given by

y[m,n] median{x[i,j],(i,j) E tu}

=

Where w represents the neighbourhood pixels.

It helps to preserve the edges and reduces noise to large
extent. After filtering the other processing is carried out.
D.CONTRAST ENHANCEMENT AND ADAPTIVE
THRESHOLDING
Now that the image is converted to gray scale and noise is
removed, to enhance the contrast, the user is allowed to stretch
it beyond its assigned values to his/her need. With this, the
slightly dark veins appear darker, hence allowing for easier
distinction between the veins and its surroundings. The range
of contrast for this image now ranges from -255 to +255 rather
than the conventional 0 to 255 pixel intensity.[5],[7].The
Figure5 : After segmentation
F. VISUALIZATION:
For an observer to view the veins more quickly, the binary
image (black and white image) is given a dash of colour. A
seed point (starting point) for the colour to start filling the
veins, depicted by a dot (pink in the image) is moved over one
vein that is convincing. The algorithm used to fill the veins is
the same as that used in the "bucket" fill algorithm. This seed
grows based on connected components hence marking the
veins distinctly as all veins are linked to the same circulatory
system.
 2013 International Conference on Signal Processing, Image Processing and Pattern Recognition [ICSIPRl 4

The test analysis of a part of the pilot test is as given in

Each of the three main processing - contrast stretching, Figure7 and TableI provided.
adaptive thresholding and the colouring of the veins are
displayed for the user to process the video to best discern the RESULT MALE FEMALE
veins with regard to ambient light and the varying anatomy FAllURE 2 1
from subject to subject.
EASY 3 3

MEDIUM 2 1
V. TEST RESULTS
HARD 1 3
The vein viewing experiment through this device was
conducted on several subjects of different sex, age, complexion TOTAL 8 8
and body type. The test result summary is provided below.
T ALEI
Example: Here is an illustration done on a subject of age
TEST ANALYSIS
20,male and having a poor visibility of veins by naked eyes.
The results were successful.
VI CONCLU SION

Our main goal of obtaining an and portable efficient vein

Figure6:The screen preview ofvein image ofa subject.
In the Figure6,we can see the preview of the laptop screen
for a subject as mentioned above. Here u can see the image
after noise removal ,then the enhanced image and finally the
veins enhanced display. The track bars are provided to adjust
the threshold for the contrast stretching to get good results for
all subjects at most of the environmental conditions. This is a
very user friendly system and gives a real-time display of the
veins. This provides instantaneous information to the doctors
and clinicians.
imaging system at a very low costs accomplished. As per the
tables provided we have done a pilot study for subjects which
includes wide range of sex, skin color and age . We have only
implemented the acquisition and segmentation of the image
which allows for future expansion. Verification performance
parameters like SNR, MSE and the pilot study success rate are
provided
A. Uncertainties
Since, we have not made an IR imaging system that is
stationary we have encountered a lot of motion artifact, we are
trying to remove the noise to a better degree so that the veins
are more c1ear and error is reduced.
B. Future work
Our system requires a better system and more
modifications are required for hardware and software to get a
clearer image, even though our success rate is good we are
trying to make it better and understand the reason behind a
failed run or an image without clear vein image.
VII ACKNOWLEDGMENT
Our gratitude to our guides Dr.K.G.Srinivasn ,Head of the
department of Computer Science Engineering , M.S.Ramaiah
Institute of technology,Bangalore
ProfG.Kumaravelu,Professor,Medical Electronics Department,
M.S.Ramaiah institute of technology, Bangalore.

4 VIII REFERENCES
[I] Adrian Kaehler Gary Bradski. Learning Opens.
2 .Male
[2] S. Crisan, l.G. Tarnovan, and T.E. Crisan. A low cost vein
o detection system using near infrared radiation. In Sensors
Male Female
Applications Symposium, 2007.SAS '07. IEEE, pages 1 {6, feb.
2007.

[3] Septimiu Crisan, loan Gavril Tarnovan, and Titus Eduard

Crisan. Radiation optimization and image processing algorithms
intheidentication of hand vein patterns. Computer Standards and
Figure7:The test analysis Interfaces, 32(3):130 { 140, 2010.XV IMEKO TC4 Symposium
on Novelties in Electrical Measurements and Instrumentation
 2013 International Conference on Signal Processing, Image Processing and Pattern Recognition [ICSIPRj 5

and X II International Workshop on ADC Modelling and [10] Combining near-infrared illuminants to optimize venous
Testing. imaging Vincent Paquita,b,c, Jeffery R. Pricea, Fabrice

[4] Medical errors in pediatric practice. M' eriaudeaub,c,Kenneth W. Tobina and Thomas L. Ferrellc

http://indianpediatrics.net/july2008/ju Iy-586-589.htm. [11] R. Sanchez-Reillo, C. Sanchez-Avilla, and A.Gonzalez-Macros,

"Biometrics identification through handgeometry
[5] T. Arici, S. Dikbas, and Y. Altunbasak. A histogram
measurements", IEEE Tans PAMI, 22(18): 1168-1171, 2000.
modification framework and its application for image contrast
enhancement. Image Processing, IEEE Transactions on, [12] http://fourier.eng.hmc.edu/e 161/1ectures/smooth sharpen/node3.
_

18(9): 1921 {1935, sept. 2009.} html

[6] H.D. Zeman, G. Lovhoiden, and C. Vrancken. The clinical [13] Toshiyuki Tanaka, Naohiko Kubo, "Biometricauthentication by

evaluation of vein contrast enhancement. In Engineering in hand vein patterns" SICE, Annual Conference in Sapporo, 249-

Medicine and Biology Society, 2004. IEMBS '04. 26th Annual 253, Aug. 2004.R. Nicole, "Title of paper with only first word

International Conference of the IEEE,volume 1, pages 1203 capitalized," J. Name Stand. Abbrev., in press.

{1206, sept. 2004.} [14] K.Z. A., Youssef, A. M., and Kadah, Y. M. "Real-Time speckle

[7] G. Lovhoiden, H. Deshmukh, C. Vrancken, Y. Zhang,H.D. reduction and coherence enhancement in ultrasound imaging via

Zeman, and D. Weinberg, "Commercialization ofVein Contrast nonlinear anisotropic diffusion". lEEE Transactions on

Enhancement," Proc. SPIE, vol. 4958,pp. 189-200, 2003 Biomedical Engineering, 49(9).

[8] Y. Sun, "Automatic identification of vessel contours in coronary [15] V. Zharov, S. Ferguson, J. Eidt, P. Howard, L. Fink, and M.

arteries in arteriograms by an adaptive tracing algorithm", IEEE Waner, "Infrared imaging of subcutaneous veins," Lasers in

Transactions on Medical lmaging, 8(1):78-88, March 1989 Surgery and Medicine 34(1), pp. 56--61, 2004.

[9] Sang-Kyun Im, Hyung-Man Park, Young-Woo Kim, Sang-Chan [16] Pascual, S., J.E., Uriarte-Antonio, J., Sanchez-Reillo, R., and

Han, Soo-Won Kim, Chul-Hee Kang andChang-Kyung Chung, Lorenz,M. Capturing Hand or Wrist Vein Images for Biometric

"An Biometric identification system by extracting hand vein Authentication Using Low-Cost Devices. In Intelligent

patterns", Journal of the Korean Physical Society, 38(3): 268- Information Hiding and Multimedia Signal Processing (IIH

272, March 2001. MSP), 2010 Sixth International Conference on (october 2010),
pp. 318