International Journal of Computer Applications (0975 – 8887)
Volume 184– No.36, November 2022
Implementing all in One Platform to Optimize the Blood
Donation Chain
Baladewa A.H.A.N.
Madhuwanthi A.V.P.
Pieris M.V.M.
Faculty of Computing
Sri Lanka Institute of Information
Technology,
Malabe, Sri Lanka
Faculty of Computing
Sri Lanka Institute of Information
Technology
Malabe, Sri Lanka
Faculty of Computing
Sri Lanka Institute of Information
Technology
Malabe, Sri Lanka
Vimukthi S.M.Y.R.
Pradeepa Bandara
Pipuni Wijesiri
Faculty of Computing
Sri Lanka Institute of Information
Technology
Malabe, Sri Lanka
Faculty of Computing
Sri Lanka Institute of Information
Technology
Malabe, Sri Lanka
Faculty of Computing
Sri Lanka Institute of Information
Technology
Malabe, Sri Lanka
ABSTRACT
Every second, someone needs donated blood or blood
products to stay alive. So, the demand has to be filled with an
efficient and optimized supply chain. Considering the
importance of utilized communication tool between the blood
donors, blood receivers, blood banks and blood donation
campaigns organizers including the health services in Sri
Lanka, a web based solution optimizing the blood donation
chain was proceeded. In the system through the machine
learning based implementations, all the users will be benefited
to aware their eligibility on donating blood before a blood
donation inputting their details while also get to know the
upcoming monthly blood requirement of the country and get
alerted on predicted requirements. Detection blood cells and
analyze percentages is another function accomplished through
the system along with a specified chat bot to assist the blood
donors and campaign organizers for the blood donation
process. The system's ultimate goal is to eliminate the timeconsuming communication barrier by connecting the
responding parties and facilitating awareness of the blood
donation procedure.
Keywords
Blood donation, Machine learning, Donors
1. INTRODUCTION
Blood donation is a process of taking blood from a person
voluntarily to be stored in a blood bank for later use in blood
transfusions [14]. Many patients with a variety of medical
issues can benefit from blood donations. Emphasizing the act
of blood donation, World Health Organization highlight the
fact that safe blood saves lives including women with
complications during pregnancy and childbirth, children with
severe anemia, often resulting from malaria or malnutrition,
accident victims and surgical and cancer patients. Human
blood is a necessary component of human existence for which
there is no alternative.Blood cannot be appraised in term of
cost, has a limited shelf-life, which must be utilized in a
relatively short period [13]. Since that, a constant need of
regular supply must be maintained.
Sri Lanka annually collects more than 350,000 blood units. To
supply the sufficient amount, it is important to make the
services available easily, effectively and efficiently. The
proposed solution is a comprehensive fully fledged system
works on optimizing the blood donation chain. The blood
donators will be registered to a system where they can be
aware about the concerns of blood donating, and get notified
in an emergency blood requirement, responding to the
monthly blood requirements and aware about the nearby
blood donations. The blood receivers will be ensured with the
ability of requesting potential blood donors, locating the
nearby places for blood transfusion, and viewing the current
availability of blood on different blood banks. The
benefactors who are willing to organize blood campaigns
doesn‟t have to waste their time on seeking the potential
places and resource suppliers. The proposed system will filter
them the best locations to organize blood donation campaigns
and recommend the responsible resource holders within a
short time. Also, the system supposed to encourage all related
parties of the system to enhance the blood donation via
sending promoting messages and notifications, fulfilling the
national cause of ensuring the sufficient, quality and adequacy
blood supply to give the life to someone.
2. LITERATURE REVIEW
Many studies have investigated people‟s attitudes toward
blood donation.Even if the primary way tomeet blood
requirement is to receive regular donations from healthy
volunteers, unfortunately, only 5% of eligible donor
population regularly donates [1]. Non-donors provide various
reasons for never donating blood, such as lack of knowledge,
fear of contracting the disease, insufficient time, not being
asked, or being medically unfit [11]. If these non-donors can
be assured that they are safe after a blood donation, whether
they are eligible to donate blood, that blood donation is a less
time-consuming process it will be a greater sustenance to
strengthen the blood supply chain [12,13].
Currently, the blood bank's working process and storage
system in Sri Lanka is mostly focused on documentations and
files. Existing method of the blood donation camp organizing
is selecting a date and place for the camp and contact the
nearest blood bank. Then organizers should meet the relevant
officers and make a request providing a letter. After that need
to wait for their acceptance and recommendations. Also, they
haven't a proper guidance and instructions system. If any
person has a problem or question regarding the blood
22
International Journal of Computer Applications (0975 – 8887)
Volume 184– No.36, November 2022
donation camp organizing or any kind of problem regarding
the process, that person need to contact the blood bank or
need to visit the blood bank to solve their problem. Data and
knowledge about blood, donors, recipients, blood donation
camps and camp organizers are stored in documents and
archives. As a result, data and information processing
becomes tough and time-consuming [1]. All blood donation
and transfusion tests and donation camp request details are
also documented on physical papers. This renders information
vulnerable to human mistakes and blunders, putting human
lives in jeopardy. Another concern with this framework is, its
lack of productivity. Organizing blood donation campaigns is
the most common way for collecting blood. Some
organizations, Non-Government Organizations and some
groups of people desire to organize blood donation
campaigns. Since it is a long process, it takes some time to
arrange and need to contact many people in the health sector
to get health aids, guidance and another parties to take the
necessary arrangements. It takes a lot of time and work to
retrieve blood, whether it's donor, receiver, or donation camp
information. Because information retrieval is such a timeconsuming procedure, hospitals have a difficult time saving
lives at critical periods. Information security and backup are
also important considerations because papers and documents
are easily stolen or misplaced. As a result, it's a shaky
structure [2].A intention of the study is to create a platform to
help and guide the campaign organizers to organize blood
donation campaigns easily through online and make a
platform to solve their problems and questions easily and
quickly regarding the camp organizing or any other problem
about blood bank or blood donation process [3]. Spent time
learning everything can about blood management systems and
processes, and used everything learned to make project the
best it can be. It was critical for us to identify vulnerabilities
in the existing systems. So that discover fixes and include
them into project [4].
counting RBCs and differentiating between normal and
pathological cells, it processed the blood smear image. They
first extracted the WBCs from the image using the K-medoids
technique and separated the RBCs from the WBCs using
granulometric analysis. Next, they labeled the image and used
a circular Hough transform to count the number of cells
(CHT). For the purpose of counting RBCs on the greyscale
image, Sarrafzadeh et alcirclet .'s transform was proposed The
circlet transform suggested by [8] can be used to count RBCs
on a greyscale image. For the objectives of counting and
identification, they adopted the iterative soft-thresholding
method. Kaur and al. suggested a technique to automatically
count platelets by applying a CHT to a picture of tiny blood
cells When using the characteristics of platelets' size and
shape from the CHT in the counting method. Cruz et al.
provided a method for processing images. a method for
counting blood cells. Color, Saturation, and Value were
employed. the associated component labeling for the
thresholding method blood cell counts and identification. [9]
Acharjee et al. using a Hough transform to suggest a semiautomated procedure. Change to count RBC by spotting their
biconcave and oval shapes. shape. An automated counting
approach was developed by Lou et al. RBCs using support
vector machines and spectral angle imaging (SVM). Using a
convolutional neural network, Zhao et al. [10] proposed an
automatic identification and classification system for WBCs
(CNN). They first identified WBCs from the microscopic
images, and then they used CNN to identify different types of
WBCs. Five different types of WBCs were categorized by
Habib Zadeh et al. They employed two distinct SVMs and one
CNN classifier, totaling three classifiers.
Blood Transfusion is a important component of any health
care industry. The timely availability of safe blood and blood
stocks is essential in health facilities where transfusion is
performed [1].There are numerous existing Web-based and
Android-based systems that record donor data and offer a
system that helps other blood banks in emergency. This
system differs from existing systems due to future scope or
view which reduces the risk of null storage at crucial times
[5]. The way that this system predicts future blood
requirements is one way in which it differs from the current
system. By maintaining and analyzing the saved data, this
system assists in forecasting the quantity of blood donations
that will be needed in the future [6]. This analysis and
management of data is lacking in the existing methods, so the
proposed methodology overcomes it with the use of machine
learning algorithms. The output is helps us understand the
future requirements. The need and the actual number of
donations are not met in country. The level of knowledge,
attitude and awareness difficulties may be the main
contributing factors. The current study also aims to make
people aware of the blood donations across the county.
Considering the current existing monthly blood donation
count with the future requirement, encouraging blood donors
spread across the country at emergency situations through a
centralized system by SMS or Email distributed through the
system is also projected [7].
Donating blood is a noble act that saves lives. Yet the health
care services of low-income countries like Sri Lanka still face
the problem of collecting sufficient blood compared to the
actual needs. So, A feature to identify the probability of a
certain person to donate blood is planned. The perspectives of
the donors and non-donors on donating blood will be
gathered, the data will be clustered based on age, health
compatibility, willingness and etc. feeding the collected data
set, a module will be trained to generate an algorithm to find
out the probability to donate blood. They will be encouraged
emphasizing the requirement of donating blood through the
system generated notifications and SMS reminders.
When counting blood cells automatically, there are often two
basic ways used. The two methods are machine learning and
image processing. An image processing method for RBC
count was proposed by Acharya and Kumar. As well as
3. PROPOSED SYSTEM
3.1 Predicting the person's probability of
donating blood and sending related
notifications on blood donation
3.2 Helping and guiding the blood donors
and campaign organizers for the blood
donation process through a chatbot
Many clinics and hospitals constantly require blood for
several purposes. Most people haven't a proper idea about the
blood donation process. So they have many questions about
this process. If they need to know more details about the
blood donation process or blood donation camp organization
process, they need to contact a doctor or blood bank. This
system intends to give information and guidance to blood
donors and camp organizers for the blood donation and camp
organizing processes through a chatbot. Users can know about
the nearest blood bank, contact details of all the blood banks
in the country, details about the blood donation process,
details about the blood donation camp organizing process,
available dates for the blood donation camps, limitations of
23
International Journal of Computer Applications (0975 – 8887)
Volume 184– No.36, November 2022
the blood donation process, information about blood types and
many more information about the blood donations. Also,
organizing blood donation campaigns is the most common
way of collecting blood. Some organizations, NGOs and some
groups of people desire to organize blood donation
campaigns. Since it is a long process, it takes some time to
arrange and needs to contact many people in the health sector
to get health aid and other parties to take the necessary
arrangements. Through the system, a solution is proposed to
overcome the difficulties faced when organizing the blood
donation campaigns via implementing an online process to
organize blood donation campaigns easily gaining information
through a chatbot on potential resource holders, places,
processes, limitations, etc.
This Chatbot can be used as a conversational chatbot. Users
can ask questions and the chatbot gives the answer. Natural
Language Processing (NLP) and Convolutional Neural
Networks (CNN) are the main methodologies used for the
chatbot. When the user adds a question, the chatbot auto
generating the answer using NLP and CNN. PyCharm is the
IDE that is used for the chatbot. PyTorch is also used for this
chatbot. PyTorch is an open-source machine learning (ML)
framework based on the Python programming language and
the Torch library. It is one of the preferred platforms for deep
learning research. The framework is built to speed up the
process between research prototyping and deployment.
Python language is used for chatbot development. When
having a good interaction with the chatbot, can get better
precision. Including deep learning, natural language
processing, and machine learning (ML) algorithms, and it
requires a huge amount of data to give accurate results. So
used a huge dataset for this chatbot.
3.3 Predicting upcoming monthly blood
requirement and encouraging people to
donate blood according to the predicted
requirements
Blood is a limited resource and a delicate product with a short
life cycle, therefore demand and supply are unstable. This
system intends to computerize in order to predict upcoming
monthly blood requirement. A module will be trained using
machine learning concepts. For this, past years data related to
monthly wise blood requirement will be gathered. Using the
past data (Date, Blood donation count) as the data set a model
will be trained. After that Past data, current data, and
predictions of future data will be shown in a graph. The need
and the actual number of donations are not met in country.
The level of knowledge, attitude and awareness difficulties
may be the main contributing factors. This system aims to
make people aware of the blood donations according to the
predicted requirements.
A machine learning algorithm is used to implement it in the
datasets. Machine learning algorithms are used to automate a
model or system, resulting in the desired output. It's done by
training the model. Using Long Short-Term Memory (LSTM),
Time series forecasting model can predict upcoming future
values based on previous sequential data. This provides
greater accuracy for forecasting and provide better decision
making for future. Time-series model is used to portrait the
monthly wise number of blood donation count. And the future
trend predictions for those measures. Past data (Date, Blood
donation count) will be needed to train and evaluate model to
predict upcoming monthly blood requirement. With this
method, the upcoming blood requirement prediction can be
done in automation way. The dataset will be containing date
and the quantity of blood donation count of every month. This
will be used to predict the further blood requirement using the
LSTM algorithm of Machine Learning. Keras and Tensorflow
libraries can be used for model training. Artificial Neural
Network will be fed with flattened set of data array and
learning model will be optimized with the loss function and
other optimization algorithms. Once the model is trained after
several epochs, model can be saved for predictions.
3.4 Detection blood cells and analyze
percentages
The field of blood cell analysis has become increasingly
important when taking into account the reach of bio
information technology. Red blood cells (RBC), white blood
cells (WBC), and platelets can now be distinguished and
counted using a variety of techniques. Currently, counting and
analyzing blood cells are done manually, which leads to
numerous human errors. Previous studies suggested a
software-web-based, economical, and effective alternative for
identifying and analyzing blood cells as a solution to this
issue. Digital image processing is used in the suggested
method. In the majority of investigations, RBCs, WBCs, and
platelets were found and extracted from the images collected
using image pre-processing and enhancement techniques like
edge detection, spatial filtering, and adaptive histogram
equalization.
Object detection aims to find the coordinates of objects in an
image and classify its category. The proposed detector regards
object detection as a regression problem in this research,
equivalent to the YOLOV5 detector's methodology. Also, the
proposed sensor that only uses the last output feature of the
backbone can converge faster and achieves promising
performance, compared to YOLOv5, which uses the feature
pyramid network to enhance performance. This proposed
detector is modified by YOLOV5.
4. RESULT AND EXPERIMENT
The prediction on eligibility of a blood donor gives 77%
accuracy by light BGM algorithm, the Gradient Boosting
machine and the regression algorithm follows the respective
second and third highest accuracy. While the LSTM (Time
series forecasting) algorithm predicts the upcoming monthly
wise blood requirements with 74% accuracy. The system
predicts detection blood cells and analyze the percentages on
70.5 % accuracy level.
The research team got feedback about this system using a few
sets of blood donors and doctors. Accordingly, 12 people
were involved in the testing phase. Overall results indicate
that most people are satisfied with the system. Also, all the
people agreed that the system is good for blood donation and
blood donation camp organizing processes, and it is better
than the traditional blood donation camp organizing process.
One of the doctors suggested it is better to add some more
details to the chatbot about the limitations used in the blood
donation process. According to that suggestion, the research
team decided to improve the data set of the chatbot. While
considering all the facts it is evident that “Blood Care” would
be a web application that can help blood donors and camp
organizers with the blood donation process.
The hospital - blood bank current system is a manually
operated device. The manual approach has issues with
keeping track of blood stock records, future blood
requirements. These records may not be preserved securely,
and records may go missing as a result of human error or any
other way. This system aims to resolve these issues faced by
24
International Journal of Computer Applications (0975 – 8887)
Volume 184– No.36, November 2022
hospitals. So, this system shifting their traditional processes
into digital technology which would not only provide a highly
efficient method capable of resolving huge complexities but
also the proper analysis of enormous data to achieve
predictions which would help us to prevent future anomalies
and medical emergencies occurring due to poor traditional
data management and processing techniques. These prediction
results would help to maintain a proper track record of future
blood requirements.
Systems on Student‟s Learning Outcomes," SYLWAN,
2019.
[4] BORKAN, "www.blog.casper.com," 21 september
2016.
[Online].
Available:
https://blog.casper.com/meet-insomnobot-3000/.
[5] Dr.J.Shanthini,SanthoshiA,Manjula T, Pavithra R,”
Computer Aided Emergency Service System”, IJIRT,
National Conference on Networks, Intelligence and
Computing Systems,March 2017.
[6] Ramesh S Patil, Sagar Mhetre, Madhavi Rayate, A
GafoorKarache,”Analysis of Blood Donor Deferral
causes in Solapur district”,International Journal of
Biological & Medical Research, nt J Biol Med Res,2014
Fig 1: Chat bot result
5. CONCLUSION
The health care services of low-income countries like Sri
Lanka still struggle to collect enough blood compared to real
needs in their health care systems. Therefore, a feature to
determine the probability that a certain person will donate
blood is planned. Additionally, they will be encouraged to
donate blood by being reminded through system-generated
notifications and SMS messages. According to the feedback
of the people who were involved in the testing phase, chatbot
and online blood donation camp organizing platform are
important for the users and it is better than the traditional
process. It can be used easily by people also who haven't a
good idea about the blood donation process. Evaluated the
traditional procedures used by blood banks for storage of
blood donation records. Based on this, a more effective, high
processing, and large amounts of data, while removing the
complexity and limitations of the present technology. This
proposed system would assist in comprehending effective data
management techniques for blood banks. And also, this
system aim to spice up cell detection accuracy, that's achieved
by adding channel attention and spatial attention mechanisms
to the feature extraction network.
6. ACKNOWLEDGMENTS
Grateful to the supervisor and co-supervisor for their immense
guidance throughout the research work. Special thanks to Sri
Lanka Institute of Information Technology for giving us the
opportunity to carry a research project which helped us to
refresh all the concepts and technologies that learned
throughout the degree. The support received from all parties is
highly appreciated.
7. REFERENCES
[1] National
Blood
Transfusion
Service
:
2022.[Online].Available: http://www.nbts.health.gov.lk/
[2] C. J. N. J. S. S. Divya Madhu, "A novel approach for
medica assistance
using
trained
chatbot,"
in
International Conference on Inventive Communication
andComputational Technologies (ICICCT), 2017.
[7] Javed Akhtar Khan and M.R. Alony ,”A New Concept of
Blood Bank Management System using Cloud
Computing for Rural Area (INDIA)”, International
Journal of Electrical, Electronics and Computer
Engineering,2015
[8] Sarrafzadeh, O., Dehnavi, A.M., Rabbani, H., et al.,:
„Circlet based framework for red blood cells
segmentation and counting‟. IEEE Workshop on Signal
Processing Systems, December 2015
[9] Acharjee, S., Chakrabartty, S., Alam, M.I., et al.,: „A
semiautomated approach using GUI for the detection of
red blood cells‟. Proc. Int. Conf. on Electrical,
Electronics, and Optimization Techniques, 2016, pp.
525– 529
[10] Zhao, J., Zhang, M., Zhou, Z., et al.,:„Automation
detection and classification of leukocytes using
convolutional neural networks‟, Med. Biol. Eng.
Comput., 2017, 55, (8), pp. 1287– 1301
[11] R.s. Ali, T.F.Hafez, A.B.Ali and N.Abd-Alsabour,
“Blood Bag: A web application to manage allblood
donation and transfution process,”2017 International
Conference on Wireless Communications, Signal
Processing and Networking(WiSPNET),2017,pp.21252130,doi: 10.1109/WiSPNET.2017.83001
[12] Domingos DC, Lima LF, Messias TF, Feijo JV, Diniz
AA, Soares HB. Blood hero: An application for
encouraging the blood donation by applying
gamification. Annu Int Conf IEEE Eng Med Biol Soc.
2016
Aug;2016:5624-5627.
doi:
10.1109/EMBC.2016.7592002. PMID: 28269530.
[13] P. KIRCI, S. AKTAS and B. SEVINC, "Analyzing
Blood Donation probabilities and number of possible
donors," 2020 International Congress on HumanComputer Interaction, Optimization and Robotic
Applications
(HORA),
2020,
pp.
1-4,
doi:
10.1109/HORA49412.2020.9152872.
[14] W. B. Zulfikar, Y. A. Gerhana and A. F. Rahmania, "An
Approach to Classify Eligibility Blood Donors Using
Decision Tree and Naive Bayes Classifier," 2018 6th
International Conference on Cyber and IT Service
Management (CITSM), 2018, pp. 1-5, doi:
10.1109/CITSM.2018.8674353.
[3] S. A. N. H. HameedullahKazi, "Effect of Chatbot
IJCATM : www.ijcaonline.org
25