International Journal of Innovative Technology and Exploring Engineering (IJITEE)

ISSN: 2278-3075, Volume-8 Issue-6, April 2019

Vein Detection System Using OpenCV with

Patient Database System
Dhanamjayulu C, Shubhangi Mittal, Biswamoy Pattnaik, Sreya Dey, Maanik Marathe, Ashutosh
Joshi, Arunkumar G

Abstract: Non-invasive vein detection for intravenous (IV) A critical problem which is encountered by the doctors and
procedurescan be carried out using infrared (IR) raysfor the medical practitioners is the visualization of patient’s veins for
purpose of illuminating a region, and then using an infrared
injecting medication and various other purposes, especially
camera for observing it. The image is processed using different
techniques on the OpenCV software platform. We further propose in case of emergency situations like trauma patients, where
the creation of a centralized database system for the management time is of essence.Medical treatments and procedures get
of patient data, which will securely store the previously captured delayed if the location of veins is improper and unreachable
images of the patient’s venous system for the healthcare due to patient's own characteristics that, in some cases, make
professional’s records, thereby increasing the staff productivity visualization or palpation of the veins difficult. More
and patient satisfaction as well as bettering overall clinical
difficulties arise in case of obese or dark toned patients as
workflow.
well as with children, especially new-born where finding
Index Terms: infrared camera, noninvasive vein detection, veins might be troublesome. For patients with bruises or
OpenCV, patient database management system burns, in which cases bruises are present on the region of
interest, the quick detection of veins is very critical and any
I. INTRODUCTION delay can cause critical issues. Redundant vein puncturing
carried out by the medical practitioners, can cause bruising of
In the field of biomedical instrumentation, current research
skin, inflammation, irritation and bleeding and therefore, the
being done is yielding efficient but high cost results. The
process of venipuncture represents an aggression for the
utilization of IR imaging procedure and imaging in broad
patient. As such, a device is required to provide the accurate
regions is a moderately less investigated territory which
position of the vein during the processof blood withdrawal,
guarantees proper results at lower costs. IR imaging can be
transfusion or donation, intravenous injection etc. where, the
carried out under two subcategories – Near IR (700-1300nm),
precise region needs to be distinguished.
which easily helps us capture more and better information,
[1] discussed using infrared light to facilitate detection of
and Far IR (8-14 µm), which gives us lower quality images.
veins in a procedure consisting of illuminating the skin with
NIR penetrates up to 3mm deep into the human tissue. The
infrared light, capture the video of the reflected image of the
basic principle of operation which is explored in this paper
veins and performing the processing of said image using
for vein detection using NIR imaging is that, the
MATLAB algorithms. [1] proposed a low-cost prototype
deoxy-hemoglobin present in the blood vessels of the human
with resources accessible at the local level that allowed
body absorbs more infrared radiation than the neighboring
obtaining venous pattern on dorsal face of the hand. An array
tissues. The veins containing deoxyhemoglobin seem darker
of NIR LEDs is made to illuminate the forearm and the image
on absorbing the infra-red light, so they become visible and
of the arm is captured by camera. The captured vein image is
the image of veins gets captured by the camera.
then is passed through levels of processing using Open CV
and image enhancement techniques are used to obtain a clear
image of the veins. The processed image of venous pattern is
available for use simultaneously, and is also uploaded to a
centralized database accessible to doctors and medical
Revised Manuscript Received on December 22, 2018.
practitioners of the hospital, along with the patient
Dhanamjayulu C, Department of Electrical & Electronics information, i.e., patient name, age, weight, as well as current
Engineering, Vellore Institute of Technology, Vellore, India body temperature and pulse rate. The database maintained is
Shubhangi Mittal, Department of Electronic and Information
Engineering, Vellore Institute of Technology, Vellore, India.
for future reference and use. The central and secure database
Biswamoy Pattnaik Department of Electronic and Information is especially useful in emergency situations where the time
Engineering, Vellore Institute of Technology, Vellore, India. for venous imaging is not available, instead the image present
Sreya Dey, Department of Electronic and Information Engineering, in the database is accessed and used. Hence, in this paper we
Vellore Institute of Technology, Vellore, India.
Maanik Marathe, Department of Electronic and Information have proposed an effective system of medical service that is
Engineering, Vellore Institute of Technology, Vellore, India. accessible, cost effective and available in case of
Ashutosh Joshi, Department of Electronic and Information emergencies.
Engineering, Vellore Institute of Technology, Vellore, India.
Arunkumar G, Department of Electrical & Electronics Engineering,
Vellore Institute of Technology, Vellore, India

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: F3945048619/19©BEIESP 1485 & Sciences Publication
 Vein Detection System Using OpenCV with Patient Database System

II. LITERATURE REVIEW radiation on to the affected body region. We can pick up the
difference between the vein to be detected and the
Much research is being conducted in the area of healthcare,
with an aim to make the practices more efficient, and increase background tissues, by capturing and processing the
the comfort level of patients while accessing medical difference in absorption and reflection of the near
services. In this section, some of the previous work in the area infraredlight in the image.
of venous imaging has been discussed which provided key B. Raspberry Pi
motivation and insight in the design and methodology
Raspberry Pi is a single-board computer used for electronics
proposed in this paper.
education and rapid prototyping. It has grown rapidly in
The measurements taken through the capturing of the image
international communities and gained importance for use in
can have an increase in inaccuracy due to physical factors
production in schools, universities and other institutions as
such as skin tone or hair on the skin. Vincent Paquita et al. [2]
well. In this model, a Raspberry Pi unit has been used as the
performed experiments to determine the best NIR
main processor for the operations-**/* required.
wavelengths to optimize vein contrast for removing the
issues caused by these. They employed an array of NIR LEDs C. The NoIR Raspberry Pi camera
and experimented with different NIR illuminations to find the Raspberry Pi camera is an image sensor of high quality
best combination of illuminant radiation by performing designed as an add-on device for use with Raspberry Pi. The
two-class linear discriminant analysis (LDA). In other NoIR Raspberry Pi camerais a similar sensor to the basic RPi
research, Koushik Kumar Nundy [3] described a vein camera, but with a modification, namely the removal of the
detection system using common VGA quality camera phones IR cut filter, allowing IR radiation to be detected by the
in order to take infra-red images., whose setup is easy having sensor. The v2 Pi NoIR has a Sony IMX219 8-megapixel
minimal cost, and the image quality is based on number and sensor and can be used for a variety of applications where the
the relative positioning of the NIRLEDs, height of system, visible spectrum is insufficient or unusable. In this model, the
and angle of inclination of LEDs. Naoto Miura et al [4] RPi NoIR camera is used to capture the IR-frequency image
proposed a method in which vein in hand is captured with of the body region under examination.
CCD with an attached IR filter to capture images in which
vein appears darker and used an attached IR filter to capture D. CLAHE algorithm
images in which vein appears darker for stable and unique Contrast Limited Adaptive Histogram Equalization, known
biometric features with automated image processing, noise popularly as CLAHE algorithm in image processing is a
reduction and normalization, pattern extraction and pattern method used for enhancement of the visibility of minute
matching algorithms. Li Xueyan et al [5] have also proposed details of an image by increasing the contrast between local
a system of biometric identification using palm vein in which regions. In this model, the algorithm is applied for helping in
the infrared palm images containing the palm vein clearly demarcating the veins from the background tissue
information are used for biometric identification. Subhasis regions as captured in the image of the patient's affected body
Chaudhuri [6] have applied image processing algorithm region.
feature extraction based on the optical and spatial properties
E. HTML&CSS
of objects for blood vessels having poor local contrast in
retinal images. Manjiree S. Waikar et al [7] have HTML which stands for Hyper Text Markup Language
implemented infrared vein detection using infrared LED and describes the structure of web pages and web applications
matching system for biometric identification using different using standard markup language. Along with Java Scripts and
algorithms in MATLAB. Marios Vlachos et al [8] have Cascading Style Sheets HTML forms the basis of creating
proposed a fully automatic method for finger vein pattern websites. HTML constructs, images and other objects
recognition using edge suppression process in which the vein embedded into web pages are the basic building blocks of
patterns are extensively enhanced. Simon Juric et al [9] have web development.
described in their paper low-cost solution for vein detection F. Firebase
using near-infrared spectroscopy in which on the vein image The Firebase Realtime Database is a cloud-hosted NoSQL
local dark lines are identified, and line tracking is executed by database that can be used to store and sync real-time data
moving along the lines, pixel by pixel. Christopher A. Mela et enabling access to the stored data from any device for better
al [10] have proposed a vein imaging system that combines coordination between users.
the reflectance mode visible spectrum images) with
transmission mode near-infrared (NIR) images in real time G. Pulse sensor
Pulse Sensor is a plug-and-play sensor used as a
III. TECHNICAL BACKGROUND non-invasive heart rate detection device. It uses infrared
In this section we have provided a description of technical radiation to measure the change in volume of blood vessels
backgrounds that will help to provide a better understanding caused due to the heartbeat.
of our proposed system.
IV. OVERVIEW OF DESIGN AND WORKING
A. Image and patient data acquisition
A. NIR LED
The acquisition system
NIR LED is a Light Emitting Diode (LED) that emits
consists of the Raspberry Pi
radiation in the near infraredzone of the electromagnetic
Camera and an arrangement
spectrum. In this model, an NIR LED array projects NIR

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: F3945048619/19©BEIESP 1486 & Sciences Publication
 International Journal of Innovative Technology and Exploring Engineering (IJITEE)
ISSN: 2278-3075, Volume-8 Issue-6, April 2019

of NIR LEDs for illuminating the region of interest (here, the histogram, threshold value of the pixel is found which is used
forearm). There is a total of twenty NIR LEDs mounted in the to distinguish between the background image and the veins.
form of a matrix and the camera is interfaced with the Division of the main image into multiple sub- image is
Raspberry Pi unit. The image captured by the camera is then performed, where sub-image contains both the veins and
processed, a description of which is provided in the next background image. For the veins i.e. the foreground image all
section. A DHT11 and a pulse sensor are also incorporated pixel values are set above the threshold value whereas for the
which collect the body temperature and the pulse of the background image the pixel value is less than the threshold
patient. This information along with vein image is displayed value. This provides effective segmentation of the image. For
to the medical personnel on site as well as transferred to the further removal of the noise and dark regions of the
central database accessible to authorized medical background image, a morphological close operation is
practitioners offsite. performed and along with that region growing is applied. In
skeletonization, the detection of the vein’s median axis is
performed. Basically, skeletonization is performed to reduce
the thickness of all lines to that of a single pixel. Finally,
segmentation is performed due to which veins are segmented
from the background image and are coated over the original
image. But to perform the pixel- based segmentation, the
region of interest should be within the threshold.

Fig. 1. The Raspberry pi camera setup used for

acquisition of vein image.

Fig. 3. Vein image before image processing is

performed.

Fig. 2. The array of NIR LEDs used to obtain distinct

image of veins
B. Image processing
The 24-bit image which is acquired using the Raspberry pi
camera in which the veins appear darker than the background
image is first converted to grayscale. In order to get a still
Fig. 4. Vein image after image processing
image, the image of the vein is taken in the frame form. Out
of the three colors red, blue & green, green color decreases
the background brightness & increases the brightness of
veins to provide a better view of the acquired image. Noise in
the image needs to be reduced, and for that, first the gray
scale image is blurred implementing the Gaussian blur
technique. On the Gaussian blurred output, Histogram
Equalization is performed in order to have an equally
distributed output image along the range of the intensity of
pixels.Using this technique, the contrast of the image is
increased and all pixels are given equal distribution between
the values of 0 to 255, as a result, veins appear darkened in Fig. 5. Vein image after image processing
comparisonwith the background. The technique of median
filtering is implemented to further remove the noise. In this
technique, median of neighborhood values replaces the pixel
value. Based on the neighborhood pixel value of the

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: F3945048619/19©BEIESP 1487 & Sciences Publication
 Vein Detection System Using OpenCV with Patient Database System

C. Centralized database for patient information 9. Simon Juric, Vojko Flis, Matjaz Debevc, Andreas Holzinger, and
Borut Zalik “Towards a Low-Cost Mobile Subcutaneous Vein
We propose in our system, the construction and maintenance Detection Solution Using Near-Infrared Spectroscopy”
of a centralized medical records database. The data is 10. Christopher A. Mela, David P.Lemmer, Forrest Sheng Bao “Real-time
accessible to only the registered medical practitioners and in dual-modal vein imaging system”
11. Shi Zhao, Yi-Ding Wang, Yun-Hong Wang - “Biometric identification
certain cases other authorized personnel like emergency based on low-quality hand vein pattern images” Seventh International
responders etc. A sign-inauthentication option is given to Conference on Machine Learning and Cybernetics, Kunming, 12-15
prevent unauthorized entry and use. Along with the captured July 2008.
12. Marios Vlachos and EvangelosDermatas “Vein Segmentation in
and processed vein image, the pulse rate and body Infrared Images Using Compound Enhancing and Crisp Clustering”
temperature of the patient are also uploaded to the database.
The main use of the central database is for patients who need AUTHORS PROFILE
to undergo medical procedures requiring vein detection on a
Dhanamjayulu C is a faculty and a member of
regular or frequent basis. For, these patients the previously control and automation department in School of
captured image of the vein detection system can be reused to electrical engineering, Vellore Institute of
save time and resources. We have also implemented pulse Technology, Vellore, India. He has received the PhD
degree in electronics engineering from Vellore
rate and body temperature detection so that a complete record Institute of Technology, Vellore, India, the M.Tech in
of the patients’ current state can be maintained in the control and instrumentation systems from Indian
database for future reference. Institute of Technology, Madras, India and the B.Tech
degree in electronics and communication engineering
from JNTU University, Hyderabad, India. He is also
V. CONCLUSION an Assistant Professor Senior with the School of Electrical Engineering,
Vellore Institute of Technology, Vellore, India. Since 2010, he has been a
Near Infra-Red imaging has been used to acquire the vein Assistant professor senior at Vellore Institute of Technology, Vellore, India.
image as it penetrates up to deep enough into the human His research activities are in the area of multilevel inverters, power
tissue to reach the surface blood vessels.The vein detection converters, and active power filters and power quality.
system can alternatively be used with real time video
capturing and processing instead of image capture. The vein Shubhangi Mittal is a student of B.tech EIE
detection system proposed in the paper was tested on department in School of electrical engineering,
Vellore Institute of Technology, Vellore, India.
different subjectshaving varying features and skin
complexions. It was found to be effectivein acquiring and
processing the captured vein image of most of the subjects.
Slight difficulty in detection occurred in case of subjects Biswamoy Pattnaik is a student of B.tech EIE
having darker skin tone and for subjects having more body department in School of electrical engineering,
hair. We also tested out the system with different number of Vellore Institute of Technology, Vellore, India.
NIR LEDs to get the minimum number of LEDs that can be
used to get a proper vein image. The obtained image was
processed using the CLAHE algorithm to create greater
Sreya Dey is a student of B.tech EIE department in
contrast between the vein and the surrounding tissues and School of electrical engineering, Vellore Institute of
enhance the veins. Additionally, DHT11 and pulse sensors Technology, Vellore, India.
were used to measure the body temperature and pulse of the
patient. All the acquired data is maintained in a centralized
patient database system. The system is user friendly not only
for the doctors and medical professionals but also for the
patients.The future work involves further processing of the Maanik Marathe is a student of B.tech EIE
department in School of electrical engineering,
vein image so as to get a clear image which can be used for Vellore Institute of Technology, Vellore, India.
further applications.

REFERENCES
1. A. Marcotti, M. B. Hidalgo and L. Mathé, “Non-Invasive Vein Ashutosh Joshi is a student of B.tech EIE
Detection Method Using Infrared Light” IEEE Latin America department in School of electrical engineering,
Transactions, Vol. 11, No. 1, Feb. 2013 Vellore Institute of Technology, Vellore, India.
2. Vincent Paquita, Jeffery R. Pricea, Fabrice M´eriaudeau, Kenneth W.
Tobina and Thomas L. Ferrellc “Combining near-infrared illuminants
to optimize venous imaging”
3. Koushik Kumar Nundy - “A low cost vein detection system using
integrable mobile camera devices”
Arunkumar G has received the Bachelor Degree
4. Naoto Miura, Akio Nagasaka, TakafumiMiyatake, “Feature extraction
in Electrical and Electronics Engineering in 2002
of finger-vein patterns based on repeated line tracking and its
and Master Degree (Power Electronics) in 2005.
application”
He completed his Ph.D. Degree at Anna
5. Li Xueyan and Guo Shuxu “The Fourth Biometric - Vein
university, Tamil nadu, India on 2015. Since
Recognition”
2015, he is working as an Associate Professor in
6. Subhasis Chaudhuri, Shankar Chatterjee, Norman Katz, Mark Nelson,
the School of Electrical Engineering, VIT
and Michael Goldbaum “Detection of Blood Vessels in Retinal Images
University, Vellore.
Using Two-Dimensional Matched Filters”
7. Manjiree S. Waikar, Dr. S. R. gengaje “ Infrared Vein Detection
System for Person Identification – An Image Processing Approach”
8. Marios Vlachos and Evangelos Dermatas “Vein Segmentation in
Infrared Images Using Compound Enhancing and Crisp Clustering”

Published By:
Blue Eyes Intelligence Engineering
Retrieval Number: F3945048619/19©BEIESP 1488 & Sciences Publication