healthcare

Article
IoT-Based Medical Image Monitoring System Using HL7 in a
Hospital Database
Md. Harun-Ar-Rashid 1,2 , Oindrila Chowdhury 3 , Muhammad Minoar Hossain 1 ,
Mohammad Motiur Rahman 1 , Ghulam Muhammad 4 , Salman A. AlQahtani 4 , Mubarak Alrashoud 5 ,
Abdulsalam Yassine 6 and M. Shamim Hossain 5, *

1 Department of Computer Science and Engineering, Mawlana Bhashani Science and Technology University,
Tangail 1902, Bangladesh
2 Faculty Member, Department of Computer Science and Engineering, Uttara University,
Dhaka 1230, Bangladesh
3 Department of Computer Science and Engineering, American International University-Bangladesh (AIUB),
Dhaka 1229, Bangladesh
4 Department of Computer Engineering, College of Computer and Information Sciences, King Saud University,
Riyadh 11543, Saudi Arabia
5 Department of Software Engineering, College of Computer and Information Sciences, King Saud University,
Riyadh 11543, Saudi Arabia
6 Department of Software Engineering, Lakehead University, 955 Oliver Road,
Thunder Bay, ON P7B 5E1, Canada
* Correspondence: mshossain@[Link]

Abstract: In recent years, the healthcare system, along with the technology that surrounds it, has
become a sector in much need of development. It has already improved in a wide range of areas
thanks to significant and continuous research into the practical implications of biomedical and
telemedicine studies. To ensure the continuing technological improvement of hospitals, physicians
now also must properly maintain and manage large volumes of patient data. Transferring large
amounts of data such as images to IoT servers based on machine-to-machine communication is
Citation: Harun-Ar-Rashid, M.;
difficult and time consuming over MQTT and MLLP protocols, and since IoT brokers only handle
Chowdhury, O.; Hossain, M.M.;
a limited number of bytes of data, such protocols can only transfer patient information and other
Rahman, M.M.; Muhammad, G.;
AlQahtani, S.A.; Alrashoud, M.;
text data. It is more difficult to handle the monitoring of ultrasound, MRI, or CT image data via
Yassine, A.; Hossain, M.S. IoT-Based IoT. To address this problem, this study proposes a model in which the system displays images as
Medical Image Monitoring System well as patient data on an IoT dashboard. A Raspberry Pi processes HL7 messages received from
Using HL7 in a Hospital Database. medical devices like an ultrasound machine (ULSM) and extracts only the image data for transfer
Healthcare 2023, 11, 139. https:// to an FTP server. The Raspberry Pi 3 (RSPI3) forwards the patient information along with a unique
[Link]/10.3390/healthcare11010139 encrypted image data link from the FTP server to the IoT server. We have implemented an authentic
Academic Editors: Ismaeel
and NS3-based simulation environment to monitor real-time ultrasound image data on the IoT server
Al Ridhawi and Ali Karime and have analyzed the system performance, which has been impressive. This method will enrich the
telemedicine facilities both for patients and physicians by assisting with overall monitoring of data.
Received: 10 November 2022
Revised: 19 December 2022
Keywords: Internet of Things (IoT); ULSM (Ultrasound Machine); RSPI3 (Raspberry Pi 3); Health
Accepted: 19 December 2022
Level Seven (HL7); Minimal Low Layer Protocol (MLLP); Message Queuing Telemetry Transport
Published: 1 January 2023
(MQTT); Globally Unique Identifier (GUID)

Copyright: © 2023 by the authors.

Licensee MDPI, Basel, Switzerland. 1. Introduction
This article is an open access article Since the Internet of Things (IoT) has a significant role in a variety of areas, including
distributed under the terms and IoT-based healthcare, intelligent buildings, and smart monitoring [1], the evolution of smart
conditions of the Creative Commons medical sensors, gadgets, cloud computing, and healthcare technology is attracting major
Attribution (CC BY) license (https://
interest from academics and the healthcare business. With the introduction of real-time
[Link]/licenses/by/
analytics in healthcare, we see technology exceeding the limits of what we can currently
4.0/).

Healthcare 2023, 11, 139. [Link] [Link]

Healthcare 2023, 11, 139 2 of 24

achieve in this sector. IoT has been recognized as being one of the most significant research
topics in the field of medicine, particularly in image processing. Furthermore, patient
engagement and satisfaction have also been enhanced, and interactions with doctors
have become easier and more efficient. Remote monitoring of a patient’s health has
also shown to lead to a reduction in hospitalizations and re-admissions. IoT is now
also having a huge impact on medical costs and clinical outcomes. However, medical
devices have now evolved to the point where a large amount of data are generated in the
healthcare environment daily. The advancement of medical technologies such as magnetic
resonance imaging (MRI), computed tomography (CT), and ultrasound imaging generates
vast amounts of data on a daily basis, and the data collected from these devices include
multiple dimensions and factors [2]. The dimensionality of medical image databases is
rising exponentially, making it difficult to manage file systems because of the rising amount
of data stored due to the growth of medical databases, so the handling of medical data has
become a top priority for healthcare providers [3].
A few research projects have created a method for transmitting real-time ultrasound
data via the internet. A smart handset or computer displays the ultrasound image as
well as other ultrasound data. For detecting kidney abnormalities, a portable ultrasound
imaging system based on FPGA and IoT is also being developed [4]. A study which
presents an algorithm for detecting and classifying renal abnormalities, uses LUT (Look
Up Tables) and SVM (Support Vector Machines) with MLP (Multi-Layer Perceptron) based
on an IoT e-health architecture [5]. Another piece of research proposes an electronic
healthcare method based on IoT. It has now become possible to transfer real-time ultrasound
video over the internet [6]. Another study has provided three types of modifications: a
modified multipath routing protocol, an NS3-program used to simulate a real-world
network to improve the modification, and a system that included the implementation of
some algorithms such as EIGRP (Enhanced Interior Gateway Routing Protocol). The goal is
to gather information about nearby nodes and add it to the routing table, then determine
the best routes while keeping the cost of the value obtained as guide information from the
router in mind. The protocol routing tables are regularly updated, resulting in topological
changes and information being updated from one router to another, resulting in additional
convergence [7]. Yet another study reviewed how to construct a new model in NS3, where
they utilized a wireless scenario for MANETs (Mobile Ad-hoc Networks) with different
parameters to analyze the behavior, and they found that the YansErrorModel had a greater
success rate in most scenarios [8]. The goal of a further piece of research was to see if it was
possible to transfer pictures of ultrasound imaging in rural/undeveloped settings. Aiming
to benefit telemedicine, they developed an Arduino-based ultrasonic developer kit [9].
Previous research states that ultrasound images are essential for medical purposes and
for physicians. As a result, it is critical that ultrasound images can be observed, and even
better if physicians can do this from remote locations with limited bandwidth. Monitoring
from remote locations is necessary, but it would be ideal to build up a real-time visualization
dashboard for all clinicians at any institution. This would bring data sources into closer
proximity to the healthcare facilities where they are analyzed [10]. The advancements in fog
computing, cloud computing, and IoT technologies are improving the healthcare industry
and might save billions of lives by enabling quicker public access to facilities and services.
Another report proposes an intelligent real-time patient monitoring system with pinpoint
accuracy that detects any irregularity in the subject’s vital data, such as temperature, and
includes a fall detection model [11]. Their main objective is to design and construct a
mobile IoT-based healthcare system that gathers patient data from different sensors and
promptly notifies both the doctor and the patient’s guardian via emails and SMS. It analyzes
the patient’s vital data remotely and detects ailments quickly [12]. Additionally, an IoT-
based human health monitoring system has also been established which measures the
user’s blood pressure, pulse, body temperature, heart rate, physiological information, and
other data about a patient’s vital signs. After collecting long-term data, factors associated
with a probable risk prediction should be investigated further in the future to expand
Healthcare 2023, 11, 139 3 of 24

the use of health monitoring systems [13]. In the future, this will provide a scientific and
practical foundation for the prevention and treatment of chronic high-risk illnesses. Another
study suggests a smart city health monitoring system. The proposed technique calculated
a spectrum by using LPCC (Licensed Professional Clinical Counselors) to discriminate
between individuals with normal and disturbed speech and detect the reason for this.
Running speech is used to test the system’s ability to grasp people’s ordinary conversations
and detect the existence of voice anomalies [14]. Perhaps this might be utilized in clinics to
provide medical practitioners with extra information on the occurrence of voice anomalies.
The recommended technology could also be placed on a soldier’s body and use the Global
Positioning System to track their health and current state. These data would be sent to
the control room via the Internet of Things. The Gustier detection system, which employs
emergency signaling sensors, would also be part of the envisioned system.

Related Works
Therefore, a model has been proposed that might be utilized to implement a low-cost
strategy for the protection of human life on the battlefield [15]. Another study shows
how IoT significantly affects the healthcare system [16]. A solution in another study
describes a patient wearing a single wristband to capture vital sign data, decreasing the
number of sensors required per individual to just one [17]. The study provides a schematic
representation for a low-cost IoT-based biomedical kit that can monitor vital signs for
persons in rural areas at a rapid rate with an accuracy of 98% and 95% for temperature
readings and pulse rate, respectively. In a further study, Musical Chairs is applied to
two well-known image recognition models, AlexNet and VGG16, and is implemented on
a Raspberry Pi network. In terms of time and energy performance, their systems were
compared to the Tegra TX2, an embedded low-power device featuring a six-core CPU and
a GPU. This configuration demonstrates how Musical Chairs enables the collaboration
of IoT devices to achieve equal real-time performance without spending the additional
costs of running a server [18]. Overall, our approach proposes helping patients deal with
challenging situations while doctors are absent. It does not help in emergency cases where
patients must be admitted to hospitals, but it does help in giving appropriate first aid
protocols for dealing with emergencies. Most people who are suffering from common
symptoms of an illness can benefit from this idea.
Medical images are typically disseminated to numerous organizations or within
organizations for correct diagnosis. Therefore, the exchange of medical data is required
to improve the quality of healthcare services. Users will be able to employ accessible
and distributed cloud computing platforms, and so IoT device connections through the
internet and advancements in cloud computing technology are required. The ability to
watch real-time picture data from an IoT server simplifies the process for both patients
and clinicians. However, due to low data transmission rates and payload size, dependable
and real-time picture transmission in IoT monitoring systems is regarded as very difficult.
IoT protocols are message-centric wire protocols developed for M2M communications that
allow the transfer of telemetry-style data in the form of messages from devices to a server
or small message brokerage over high latency or limited networks. MQTT, MLLP and
CoAP meet these requirements by utilizing tiny message sizes, message management, and
lightweight message overhead. Real-time picture data monitoring systems using MQTT
and MLLP protocols are difficult for IoT servers to handle. This study proposes a strategy
to remedy this problem.
The main contributions of this study are given below:
1. Machine-to-machine-based communication with an IoT server using HL7 mes-
sage processing.
2. Medical diagnoses image data stored in an FTP server with unique encrypted
image links.
3. A real-time image data display with patient info in an IoT server.
Healthcare 2023, 11, 139 4 of 24

The rest of the paper is organized as follows: (i) proposed methodology, with discus-
sion on how to transfer telemetry data with images; (ii) results analysis; (iii) discussion,
and (iv) conclusions.

2. Materials and Methods

Machine-to-machine-based data communication and monitoring of patient diagnostic
information over IoT servers is difficult and time consuming. To overcome the above
problem, this study proposes a new method where the following protocols and methods
are used to improve accuracy: MLLP protocol, HL7 messaging, FTP based storage, RSA
based encryption and decryption, and MQTT protocol. This study uses HL7 messaging
for machine-to-machine-based communication, and for large data storage uses an FTP
server. The MQTT protocol is used to communicate with an IoT server using RSA-based
encryption and decryption.
In this approach, our proposed model divides the entire procedure into four parts,
which are as follows:
• Ultrasound Machine to Raspberry Pi HL7 messaging process;
• Result data stored in LIS (Lab Information System);
• Raspberry Pi to FTP server image transfer;
• Transfer of files from FTP server to MLLP and MQTT protocol with the help of a
Healthcare 2023, 11, x
Raspberry Pi 3 (RSPI3). 5 of 26
The above steps are displayed as a flowchart in Figure 1.

Figure 1. Flow chart of the whole process.

Figure 1. Flow chart of the whole process.
This study obtained a variety of options from the analysis for using database plat-
forms to track the health status of patients. Comprising observations and research proce-
dures, this study provides solutions for monitoring patient data. From looking at previous
studies, this study also found several bottlenecks when monitoring image data in real time
in hospital databases.
Before being sent to cloud storage, a microprocessor processes all of the output raw
Healthcare 2023, 11, 139 5 of 24

This study obtained a variety of options from the analysis for using database platforms
to track the health status of patients. Comprising observations and research procedures,
this study provides solutions for monitoring patient data. From looking at previous studies,
this study also found several bottlenecks when monitoring image data in real time in
hospital databases.
Before being sent to cloud storage, a microprocessor processes all of the output raw
data. With the use of IoT, physicians can simply view that information. This solution heav-
ily relies on the Raspberry Pi for real-time data processing and the IoT server i.e., things-
[Link], an open-source IoT platform [19]. However, the main issue of transmitting image
data directly to the IoT server via the MQTT and MLLP protocols is handled in our pro-
vided approach with a new methodology which overcomes the image viewing problem
in real time. A custom widget in the IoT server also aids in the deployment of a machine-
Healthcare 2023, 11, x to-machine based custom dashboard with image and physician-based access. 6 ofThe
26 overall

schematic diagram is shown in Figure 2.

Figure Schematic
Figure [Link] diagram
diagram (full(full system).
system).

However,physicians
However, physicians maymay not to
not need need to consider
consider latency,bandwidth
latency, suitable suitable bandwidth
manage- man-
agement,
ment, or fog orcomputing,
fog computing, or to
or to wait forwait for clearance
clearance from others from others therapy
for precise for precise therapy and
and pre-
scriptions, if theifdiagnosed
prescriptions, ultrasound
the diagnosed images are
ultrasound processed
images and transformed
are processed with HL7
and transformed with
in such a way that only uses the patient’s sample ID for the entire communication.
HL7 in such a way that only uses the patient’s sample ID for the entire communication. Con-
verted ultrasound
Converted ultrasoundimagesimages
of patientof diagnoses will be stored
patient diagnoses willinbehospital
stored databases,
in hospital anddatabases,
precise ultrasound image conversion and sharing links to the images in
and precise ultrasound image conversion and sharing links to the images in an IoT server an IoT server will
ensure precision monitoring of patient illnesses, and thus patients will receive appropriate
will ensure precision monitoring of patient illnesses, and thus patients will receive ap-
treatment from doctors at any time. The proposed model directly acquires data from ul-
propriate treatment from doctors at any time. The proposed model directly acquires data
trasound equipment. The video is converted to photos using the most up-to-date ultra-
from ultrasound equipment. The video is converted to photos using the most up-to-date
sound machine, which is ideal for the experimental results. The transformed image is sent
ultrasound
to RSPI3 in HL7machine,
base64which
format,is along
ideal with
for the experimental
patient data. RSPI3 results. Thethe
processes transformed
image and image is
sent to RSPI3
converts it into in HL7
a jpg orbase64
png file,format, along
which will thenwith patient data.
be delivered to theRSPI3 processes
FTP server. Again,the image
and converts(IP
socket-based it into a jpgcommunication
+ PORT) or png file, which willthe
connects then be delivered
ultrasound to the
machine FTPRSPI3.
to the server. Again,
socket-based
One side is the(IP + PORT)
listener, while communication connects
the other is the sender, or the
viceultrasound
versa. In thismachine
instance,to the RSPI3.
the
One side method
proposed is the listener,
uses an while
HL7-basedthe other is the sender,
data transfer [Link] vice versa.
Ultrasound In thisoften
machines instance, the
convert ultrasound
proposed method raw usesdata to video exclusively.
an HL7-based data transfer Physicians
method. extract visuals from
Ultrasound the
machines often
video that
convert are ideal raw
ultrasound for visualizing
data to video patient information.
exclusively. This image
Physicians is transformed
extract visuals fromtothe video
base64
that areformat
idealbefore being sent topatient
for visualizing the LIS information.
computer. An RSPI3 This was
image used
is as the LIS in thisto base64
transformed
case. This
format RSPI3being
before takes sent
an HL7 message
to the and converts
LIS computer. Anthe base64
RSPI3 wasimage
usedtoasJPGtheor PNG
LIS in this case.
format using the HL7 messaging protocol. The image quality is also checked by the RSPI3.
This RSPI3 takes an HL7 message and converts the base64 image to JPG or PNG format
The patient sample ID is sent in an HL7 message with each image data (see Section 2.1.1).
using the HL7 messaging protocol. The image quality is also checked by the RSPI3. The
To assign and show the patient information through a machine-to-machine dash-
patient
board on sample
an IoTID is sent
server, theinproposed
an HL7 message with each
method utilizes image datain(see
[Link] thisSection 2.1.1).
case [19].
Here, our primary research experiment is to set up monitoring of image data collected
from hospital machines using IoT servers, with the process going via Raspberry Pi to the
IoT server. The proposed model uses ultrasound equipment for this purpose, which will
gather ultrasound images from patients and monitor the image data from the IoT server
without transferring the image files. The core benefit here is that we can easily monitor
Healthcare 2023, 11, 139 6 of 24

To assign and show the patient information through a machine-to-machine dashboard

on an IoT server, the proposed method utilizes [Link] in this case [19]. Here, our
primary research experiment is to set up monitoring of image data collected from hospital
machines using IoT servers, with the process going via Raspberry Pi to the IoT server. The
proposed model uses ultrasound equipment for this purpose, which will gather ultrasound
images from patients and monitor the image data from the IoT server without transferring
the image files. The core benefit here is that we can easily monitor image data from an
IoT server without sending image files in real time. The proposed model also transfers
telemetry data and an image-encrypted link to the IoT server.
Healthcare 2023, 11, x 7 of 26

2.1. Ultrasound Machine to Raspberry Pi HL7 Messaging Process

Ultrasound technology has advanced dramatically in the last decade, enabling tremen-
2.1. Ultrasound Machine to Raspberry Pi HL7 Messaging Process
dous Ultrasound
downsizing in ultrasound equipment design and manufacture. Presently, ultrasound
technology has advanced dramatically in the last decade, enabling tre-
technology rangesinfrom
mendous downsizing smallequipment
ultrasound machines thatand
design fitmanufacture.
in the palm of one’s
Presently, hand to high-end
ultra-
equipment that can
sound technology provide
ranges complicated
from small machines that ultrasound
fit in the palmtesting.
of one’s hand to high-
end equipment
HL7 messagesthat can are
provide
usedcomplicated
to transferultrasound testing. systems and make a sequence of
data between
HL7 messages are used to transfer data between systems and make a sequence of
segments, with individual segments or groups of segments being optional, obligatory, or
segments, with individual segments or groups of segments being optional, obligatory, or
repeated (HL7
repeated (HL7 cardinality).
cardinality). The message
The message timingindiagram
timing diagram in Figure
Figure 3 illustrates how3 the
illustrates how the
ultrasound machine
ultrasound machine (ULSM)
(ULSM) and interact
and RSPI3 RSPI3 with
interact with one
one another. Theyanother. They
interact with eachinteract with each
other on
other ondemand
demand viavia
intranet connection,
intranet based onbased
connection, order andonresult
ordermessages.
and result messages.

Figure 3.
Figure [Link]
Messagetiming diagram
timing from anfrom
diagram ultrasound machine to Raspberry
an ultrasound [Link] Raspberry Pi.
Healthcare 2023, 11, 139 7 of 24
Healthcare 2023, 11, x 8 of 26

The ULSM
The ULSM communicates
communicateswith withthe
theRSPI3
RSPI3based
basedon onmessaging
messagingprotocols
protocolswhich
whichuse
use the
following sequence:
the following sequence:
(i). ConnectionMessage
(i). Connection Message
(ii). ConnectionAcknowledgement
(ii). Connection Acknowledgement
(iii). Order
OrderMessage
Message
(iv). Order
OrderAcknowledgement
Acknowledgement
(v). Result
ResultMessage
Message
(vi). Result
ResultAcknowledgement
Acknowledgement
They check
They checkininwithwitheach
eachother
otherviavia a connection
a connection message,
message, NMDˆNO2ˆNMD_NO2,
NMD^NO2^NMD_NO2,
before
before communicating.
[Link] theconnection
connection is effective, thethe
is effective, other component
other component responds withwith
responds
an acknowledgement
acknowledgement message, message, ACKˆN02ˆACK
ACK^N02^ACK (Figure (Figure 3)3)and
andthen
thenthe
theconnection
connectionis is
made.
made. Otherwise,
Otherwise, it will itnotify
will notify the that
the user user the
thatother
the other machine
machine is down
is down or that
or that thethe connec- has
connection
tion
beenhas been The
broken. broken.RSPI3The RSPI3
sends ansends
order an order
to the ULSMto the ULSM
based on based on the connected
the connected patient order
patient
with order
that withID.
sample thatAfter
sample ID. After
acquiring theacquiring the order
order message, themessage, the ULSM
ULSM places it in places it An
a queue.
in a queue. An acknowledgement
acknowledgement message, ACK^OR02^ACK,
message, ACKˆOR02ˆACK, is returned. is returned.
The
The same
sameprocedure
procedureisisused usedtotosend
send the observation
the observation result message
result (Figure
message 4). 4).
(Figure

Figure
Figure 4.
4. HL7
HL7communication
communicationprocess,
process,ultrasound machine
ultrasound to Raspberry
machine Pi. Pi.
to Raspberry

The Minimal
The MinimalLowerLowerLayer
LayerProtocol
Protocol(MLLP)
(MLLP)and andSocket
SocketProgramming
Programmingare aretwo
twoproto-
protocols
cols
that that
mustmust be followed
be followed when when transmitting
transmitting HL7HL7 (HealthLevel
(Health Level7)7)messages
messages from
from one
one end
end point to another. The proposed model uses socket programming
point to another. The proposed model uses socket programming processes where listener processes where lis-
tener ports are open in distinct ports on both sides to accept incoming messages.
ports are open in distinct ports on both sides to accept incoming messages. Both systems are Both
systems
linked viaareintranet.
linked via intranet.
RSPI3 RSPI3antransmits
transmits order listanfrom
order thelist
LISfrom theULSM,
to the LIS to the ULSM,
which receives
which
it and stores it in a queue on the local system. ULSM, on the other hand, sends hand,
receives it and stores it in a queue on the local system. ULSM, on the other the result
Healthcare 2023, 11, x 9 of 26
sends
to the the result
RSPI3, to thereceives
which RSPI3, which
the listreceives the list
as a queue. HL7as amessages
queue. HL7 are messages are used
used to communicate
to communicate
between the twobetween
[Link] two sections.
The
The HL7
HL7message
messageisisencoded
encodedinin bytes
bytesand
andsent to the
sent other
to the sideside
other (Figure 5). 5).
(Figure

Figure 5. HL7 protocol layer-based connection.

Figure 5. HL7 protocol layer-based connection.

2.1.1. HL7 Block Formation and Messages

The HL7 block format is as follows:
Healthcare 2023, 11, 139 8 of 24

2.1.1. HL7 Block Formation and Messages

The HL7 block format is as follows:

<SB>Message<EB><CR>

Here:
<SB>—SB stands for Start Block Character, which is a 1-byte ASCII character, such as

<VT> or <0x0B>.

Message—This denotes a message with a certain number of bytes of data. It is a block

of HL7 data content.
<EB>—End Block, or EB, is a 1-byte ASCII character such as <FS> or <0x1C>.
<CR>—CB denotes a carriage return containing one byte of data. For instance, 0x0D.
There are two types of connection from the ULSM to the RSPI3 and vice versa:
(a) Active connection and (b) Passive connection.
Here, the proposed model used an active connection –
(a)Active Connection:
Connection established on both sides (i) Persistent (ii) Transient
(i) Persistent
1. On startup, try to connect to the listening or passive system’s set IP address and port.
2. If the connection fails, wait a brief time before retrying the connection. Once all
possible retries have been used up, the connection will either be deactivated at the
operator’s request, or the instrument will be switched off.
3. Maintain the connection at all times, even if there are no further messages to send.
4. Retry the connection if a disconnection is detected outside of a data exchange.
5. A message transmission fault is displayed if a disconnection is detected during a
data exchange (for example, while waiting for a response), and the ULSM stops the
HL7 connection.
The above steps are described in Algorithms 1–3 accordingly.
Algorithm 1 shows how the ULSM communicates with the RSPi3, Algorithm 2 shows
how the RSPi3 communicates with the ULSM using socket programming, and Algorithm 3
describes how the listener receives HL7 messages.

Algorithm 1: Ultrasound Machine to RSPI3 Message Transfer

1 Start
2 while UltrasoundMachineWantsToSendMessage do
3 if is ConnectionIsActive then
4 Send HL7 Message to RSPI3 [SB EB CR]
5 else
6 while!isConnectionIsActive do
isConnectionIsActive = Established A Socket or
7
TCP/ IP [IP + PORT] Connection With RSPI3
8 SLEEP(Few Seconds)
9 End

(ii) Transient
1. Try to connect only when a message is ready to be sent.
2. When a message is ready to be sent, connect to the listening or passive system’s preset
IP address and port.
3. If the connection fails, it needs to wait a few moments before attempting it again, as
in the Persistent method.
4. Only keep the connection open if there are further messages to send. Close the
connection if this is not the case.
Healthcare 2023, 11, 139 9 of 24

5. A message transmission fault is displayed if a disconnection is detected during a

data exchange (for example, while waiting for a response), and the ULSM stops the
HL7 connection.

Algorithm 2: RSPI3 to Ultrasound Machine Message Transfer

1 Start
2 while UltrasoundMachineWantsToSendMessage do
3 if isConnectionIsActive then
4 Send HL7 Message to Ultrasound Machine [SB EB CR]
5 else
6 while !isConnectionIsActive do
isConnectionIsActive = Established A Socket or
7 TCP/ IP [IP + PORT] Connection With
Ultrasound Machine
8 SLEEP(Few Seconds)
9 End

Algorithm 3: Listener
1 Start
2 while ConnectionIsOpenForListning do
Receive HL7 Message in byte code and
3
Convert it to Readable format
4 End

In HL7 communication, different kinds of messages are used, such as the Connection
Messages (see Table A1), Order Messages (Table A2), Result Messages (Table A3), and
Acknowledgement Messages (Table A4).

2.2. Result Data Extraction

The ultrasound machine sends a result message based on HL7 via a socket-based
connection. The HL7 result message contains a few segments which are MSH (Message
Header), PID (Patient information with ID), SID (Sample ID), OBR (Observation Result),
and OBX (Observation Result Data). From those segments, the proposed model extracts
that information via a segment-based data extraction model, which is a string-based
concatenation model.
After that, the extracted result data are used for further processes such as (i) storing in
the database, or (ii) data processing for FTP and IoT.

Result Data Stored in the LIS (Lab Information System)

(a) Raw data of the results are stored in the LIS (Laboratory Information System) via API
(Application Programming Interface).
(b) Extracted data are:
(i). Patient’s info
(ii). Sample ID
(iii). Extracted Result data stored in the database.
After getting the result message from the ULSM (Ultrasound Machine), the proposed
model categorizes the message into two parts: (i) saved raw result data (ii) extracted result
data. Here, the proposed model worked with (ii) the extracted result data.
Here, real-time data mean raw formatted data which have been sent by ULSM and are
stored in LIS storage, and thereafter any changes will be strictly prohibited.
The result data are transferred from the RSPI3 to the LIS via the API over the intranet
connection (Figure 6). Here, messages are stored in the main database every time the
validity of upcoming messages are checked in LIS based on the API token. Therefore, no
unwanted message will be stored in the LIS because of the API token. The raw result
data are stored in the database as a response. Extracted data from result messages are also
 Here, real-time data mean raw formatted data which have been sent by ULSM and
are stored in LIS storage, and thereafter any changes will be strictly prohibited.
The result data are transferred from the RSPI3 to the LIS via the API over the intranet
connection (Figure 6). Here, messages are stored in the main database every time the va-
Healthcare 2023, 11, 139 lidity of upcoming messages are checked in LIS based on the API token. Therefore, no24
10 of
unwanted message will be stored in the LIS because of the API token. The raw result data
are stored in the database as a response. Extracted data from result messages are also rec-
orded in the
recorded database
in the using
database thethe
using patient record
patient IDID
record and sample
and sampleID,ID,which
whichisisthen
thenutilized
utilized
for further processing.
for further processing.

Figure
Figure6.6.RSPI3
RSPI3totoLIS
LISresult
resultdata
datatransfer.
transfer.

[Link]
2.3. RaspberryPiPitotoFTPFTPServer
ServerImage
ImageTransfer
Transfer
Aftergetting
After gettingaamessage
messagefrom fromthe theultrasound
ultrasoundsourcesourceand
andstoring
storingdata
dataininthe
theLIS,
LIS,the
the
Raspberry Pi processes the HL7 messages and finds out patient
Raspberry Pi processes the HL7 messages and finds out patient information, sample ID, information, sample ID,
and doctor information using the Base64 image. The Base64 image
and doctor information using the Base64 image. The Base64 image is transferred to the is transferred to the FTP
Healthcare 2023, 11, x server
FTP database
server as a as
database formatted
a formattedimage with with
image a Globally Unique
a Globally Identifier
Unique (GUID).
Identifier The
(GUID).
12 of 26 image
The
is renamed with the GUID and sent to the FTP server using the
image is renamed with the GUID and sent to the FTP server using the FTP server user FTP server user ID and
ID
password.
and [Link] the image
After is successfully
the image transferred
is successfully to thetoFTP
transferred theserver, it responds
FTP server, with a
it responds
unique
with imageimage
a unique link using the GUID. This process is described in the following steps:
(a) Generate the GUID. link using the GUID . This process is described in the following steps:
(i). Convert
(i). Convert thethe image
image from
from Base64
Base64 to JPG, or getimage
the image fromorHL7 or theBase64
direct
(b) Transfer Base64 image data using thetoGUID
JPG, or
toget
thethe
FTP [Link] HL7 the direct
Base64 image.
image.
(c) Get Base64 image links from the FTP server.
(ii). Construct the GUID (Globally Unique Identifier) for the image.
(ii). Construct
(d) Encrypt the image thelink
GUID via(Globally
public key. Unique Identifier) for the image.
(iii). Rename this image with the GUID send it to FTP server.
(iii). Rename
(e) Transfer this imageimage
the encrypted with link
the GUID
to the send it to FTP server.
IoT server.
(iv). From the FTP server generate a unique image link and send it back to the Raspberry
(iv). From
Those Pi. imagesthe FTP server generate a unique image IDlink and sendasitper
back tofollowing
the Raspberry
Afterare also mapped
getting basedmessage
the response on sample from withFTP
the GUID
server the theRSPI3 encrypts the
steps: Pi. After getting the response message from the FTP server the RSPI3 encrypts the
image link via RSA public key and then sends the link to the IoT server (Figure 7).
image link via RSA public key and then sends the link to the IoT server (Figure 7).
Algorithm 4 is used in this process to:

Figure 7. Image transfer process from RSPI3 to FTP.

Figure 7. Image transfer process from RSPI3 to FTP.

Encryption Process
The following RSA encryption process is carried out based on the GUID of an image.
This study uses the key and messages as shown in Tables A5–A9.
Algorithm 4 describes how the Raspberry Pi sends the image to the FTP server.
Healthcare 2023, 11, 139 11 of 24

Algorithm 4 is used in this process to:

(a) Generate the GUID.
(b) Transfer Base64 image data using the GUID to the FTP server.
(c) Get Base64 image links from the FTP server.
(d) Encrypt the image link via public key.
(e) Transfer the encrypted image link to the IoT server.
Those images are also mapped based on sample ID with GUID as per the follow-
ing steps:

Encryption Process
The following RSA encryption process is carried out based on the GUID of an image.
This study uses the key and messages as shown in Tables A5–A9.
Algorithm 4 describes how the Raspberry Pi sends the image to the FTP server.

Algorithm 4: RSPI3 to IoT Server Image Transfer

Input: Image data from Ultrasound Machine
Output: boolean:true/false
1 Start
2 IoTServerResponse = false
3 while GetImageDataFromUltrasoundMachine do
4 if ImageDataIsValid then
5 ftpResponse = FTPServer(imageData)
6 if ftpResponse == IsValidImageLink then
imageLink = ftpResponse[‘imageLink’]
encryptedImageLink = ENCRYPT(imageLink, publicKey)
7
IoTServerResponse = TransferImageToIoTServer
(encryptedImageLink)
8 return IoTServerResponse
9 End

2.4. RSPI3 to IoT Server Image Transfer

The MQTT protocol is used to send the encrypted image link to the IoT server. After
producing an encrypted image link with a login and password from an FTP server, the
image link is transferred to a real-time IoT server. This phase configures the real-time
IoT server. This image is displayed in the IoT server alongside patient data. If a patient’s
ultrasound image cannot be transmitted to an IoT server, the image is re-sent. The Raspberry
Pi not only sends the encrypted image, but it also sends the patient information using the
MQTT protocol. This process is described in the following steps:
(i). Get the encrypted image link from the FTP server.
(ii). Transfer the encrypted image link to the IoT server using the IoT client username and
password.
(iii). Get the IoT response for image link transfer, if yes then step 4 else step 1.
(iv). Then, decrypt the image link and display it in the IoT server based on the client.
These steps are based on Algorithm 5.
Algorithm 5 describes the Raspberry Pi to IoT server encrypted link transfer process.
Healthcare 2023, 11, 139 12 of 24

Algorithm 5: RSPI3 to IoT Server Image Link Transfer

Input: encryptedImagelink, patientInfo encryptedImagelink is a string and patientInfo
finite set
patientInfo = {patientInfo1 , patientInfo2 , . . . , patientInfon } of JSON
Output: boolean info true or false
1 Start
2 while EncryptedImageLinkIsValid do
3 IoTServerResponse = IoTServer(encryptedImagelink,patientInfo)
4 if IotServerResponse == false then
5 IoTServer(encryptedImage,patientInfo)
6 Function IoTServer(encryptedImagelink,patientInfo):
ecryptedImageLink = DECRYPT(encryptedImagelink,privateKey)
7
isDisplayed = displayInfo(decryptedImageLink,patientInfo)
8 return isDisplayed
9 Function displayInfo(decryptedImageLink,patientInfo):
10 isDisplayed ← false
11 if decryptedImageLinkisanImage And patientInfo then
12 Display Image to assign Dashboard in IoT Server
13 isDisplayed ← false
14 return isDisplayed
15 End

3. Results and Discussion

Data collected directly from hospital experiment machines show that ultrasound
equipment is not practicable since the observed data are in a raw format and require
preprocessing as well as a LIS (lab information system). However, a large number of
devices send data to several patients at the same time. Managing and monitoring such
operations in real time is thus more challenging and time intensive. Finally, the proposed
model was able to reach an overall situation where this model could calculate data response
times based on bandwidth and data packet size. The proposed model also measured the
patient information sent, the data transmitted each time, and the total data sent each time
for the given criterion. The result was created based on the minimum time it took to send
the data to the IoT server.
The LIS and RSPI3 connection is shown in Figure 8. The patient information is sent
Healthcare 2023, 11, x using an HL7 message after extraction. They are connected using socket programming 14 of
via26
an ethernet cable.

[Link]
Figure LIStotoRSPI3
RSPI3connection.
connection.

Patient information results are shown in the IoT server where the proposed model
uses [Link] in a demo server. In this part, the proposed model needs to configure
the IoT server to display that information. In Figure 9, the thingsboard configuration is
shown. Figure 9a shows the possible thingsboard options for configuration and Figure 9b
shows a machine-to-machine based dashboard (for the proposed model only one dash-
Healthcare 2023, 11, 139 13 of 24

Patient information results are shown in the IoT server where the proposed model uses
[Link] in a demo server. In this part, the proposed model needs to configure the
IoT server to display that information. In Figure 9, the thingsboard configuration is shown.
Figure 9a shows the possible thingsboard options for configuration and Figure 9b shows a
machine-to-machine based dashboard (for the proposed model only one dashboard was
used). Figure 9c shows the most important part of the image display, a custom widget
which is customized for image viewing only. This widget is customized for image link
decryption using the RSA private key and displays the image which is linked with the
dashboard. User-based dashboard access control for the proposed model is shown in
Figure 9d, where access to specific dashboards can easily be assigned to users. A summary
of11,the
Healthcare 2023, x overall configuration of the thingsboard is explained in Table 1. 14 of 26

Table 1. Thingsboard configuration summary (based on Figure 10).

Figure Type Feature List Description

Signifies the data type, whether they
Rule Chains
are telemetry data or patient data
Assign/Monitor customer roles in
Customer Management
dashboard
The type of device that will be directly
(a) Thingsboard
Widget Assets connected to the IoT server i.e., ULSM,
Dashboard (Main)
RSPI3
Devices Shows the list of device type and
Device Profiles connections

Installs and deploys system updates

Figure 8. LIS to RSPI3OTA Updates
connection.
over-the-air (OTA) to devices
(a) Ultrasound Patient information results are shown in the IoT server where the proposed model
Action Buttons (View, Edit,
Machine-based Summary Table uses [Link] in a demo server. In this part, theOrganize
proposed and
model needs the
control to configure
data
Download,
the IoT server to display Share, Delete) In Figure 9, the thingsboard configuration is
that information.
Dashboard
shown. Figure 9a shows the possible thingsboard options for configuration and Figure 9b
shows a machine-to-machine based dashboard (for Makes sure to decrypt
the proposed the generated
model only one dash-
(c) Custom Widget
Source Code board was used).Custom
Figure Modification encrypted
9c shows the most important part oflink by using
the image RSAa private
display, custom
for image display
widget which is customized for image viewing only. key to widget
This displayispatient image
customized for data
image
link decryption using the RSA private key and displays the image which is linked with
(d) User-Based Data entry for customer info and role
Popup Module the dashboard. User-based
User Roles
dashboard access control for the proposed model is shown in
Dashboard Assign assignment
Figure 9d, where access to specific dashboards can easily be assigned to users. A summary
of the overall configuration of the thingsboard is explained in Table 1.

(a)

Figure 9. Cont.
Healthcare 2023, 11, 139 14 of 24
Healthcare 2023, 11, x 15 of 26

(b)

Healthcare 2023, 11, x 16 of 26

(c)

Devices
Shows the list of device type and connections
Device Profiles
Installs and deploys system updates over-the-air (OTA)
OTA Updates
to devices
(a) Ultrasound Action Buttons (View,
Summary
Machine-based Edit, Download, Share, Organize and control the data
Table
Dashboard Delete)
(c) Custom Widget Source Makes sure to decrypt the generated encrypted link by
Custom Modification
for image display Code using RSA private key to display patient image data
(d) User-Based Popup
User Roles Data entry for customer info and role assignment
Dashboard Assign Module
(d)
After configuring
Figure [Link] with custom(a)widgets,
Thingsboard configuration: physicians
Thingsboard dashboardcan easily
(Main); (b) see patientmachine
ultrasound
Figure 9. Thingsboard configuration:
dashboard.
(a) Thingsboard dashboard (Main); (b) ultrasound machine
information with images (c) Displaying
(Figure 10). images
On theusing
left aside
custom
of widget;
Figure(d)
10,assigning user-based
ultrasound image dashboards
data (ac-
dashboard. (c) Displaying
are displayed. images
cess control).
On the right using ainformation
site patient custom widget; (d) assigning
is shown. This is theuser-based
final output dashboards
of
(access control).
the image and patient
Table [Link] which configuration
Thingsboard is displayedsummary
on the (based
IoT server in real-time.
on Figure 10).

Figure Type Feature List Description

Signifies the data type, whether they are telemetry data
Rule Chains
or patient data
(a) Thingsboard
Widget Customer Management Assign/Monitor customer roles in dashboard
Dashboard (Main)
The type of device that will be directly connected to the
Assets
IoT server i.e., ULSM, RSPI3

Figure 10. Ultrasound image view with patient information.

Figure 10. Ultrasound image view with patient information.
After requesting patient information from the IoT server, the proposed model exam-
ines different sized images based on different bandwidths and checks the IoT server re-
sponse for displaying those images. Here, the proposed model uses 1 MB to 5 MB images
which are displayed on the IoT server depending on the bandwidth. Table 2 shows that
image display delay time varies based on image size as well as bandwidth. Lower band-
Healthcare 2023, 11, 139 15 of 24

After configuring thingsboard with custom widgets, physicians can easily see patient
information with images (Figure 10). On the left side of Figure 10, ultrasound image data
are displayed. On the right site patient information is shown. This is the final output of the
image and patient info which is displayed on the IoT server in real-time.
After requesting patient information from the IoT server, the proposed model examines
different sized images based on different bandwidths and checks the IoT server response
for displaying those images. Here, the proposed model uses 1 MB to 5 MB images which
are displayed on the IoT server depending on the bandwidth. Table 2 shows that image
display delay time varies based on image size as well as bandwidth. Lower bandwidth
with small packet size means that it takes more time to show the information on the IoT
server. On the other hand, higher bandwidth with higher packet size means it takes less
time to display (Table 2).

Table 2. Image display in IoT server by category.

Image Category Display Medical Patient

Delay
(2D) Total Image Data Size Network Image Information
Minimum (s)
per 30 s Correctly (s) Found (s)
5 > image Image < 1 MB 2G 15 10 5
5 < image < 10 1 MB < Image < 3 MB 3G 10 5 3
15 < image < 20 3 MB < Image < 5 MB 4G (LTE) 5 2 2
Broadband above
Image > 20 Image > 5 MB 3 2 2
5 Mbps

In addition to the real-time situation, the proposed model also simulates the whole
process using NS3 in two parts: (i) ULSM to hospital management system (HMS), and
(ii) HMS to FTP and IoT.
NS3 allows us to trace the node pathways, which allows us to determine the number
of data points sent or received. Trace files are created to track these activities. Its main
purpose is to mimic networks of connecting nodes and the traffic that flows between them.
To do this, NS3 offers a core abstraction of computer nodes with applications to produce
traffic as well as net devices and channels to transport the traffic.
In part (i) of the ULSM to HMS simulation, image data transfer employed a total of
10 nodes, each labelled with a number from 1 to 10. Node 1 is the source node from which
the proposed model transfers packet data using ULSM. This study employed nodes 2 to
9 as intermediate nodes, such as the RSPI3, and Node 10 as a server, receiving packets from
different nodes and responding with acknowledgement (ACK) messages (Figure 11).
In part (ii), packets are sent from HMS to FTP and from HMS to IoT (Figure 12). This
simulation utilized various nodes in this case, with node 0 serving as the source, node
1 serving as the HMS, node 2 serving as the FTP server, and node 3 serving as the IoT server.
Tables 3–5 emerged from those simulations.
 the proposed model transfers packet data using ULSM. This study employed nodes 2 to
9 as intermediate nodes, such as the RSPI3, and Node 10 as a server, receiving packets
from different nodes and responding with acknowledgement (ACK) messages (Figure 11).
In part (ii), packets are sent from HMS to FTP and from HMS to IoT (Figure 12). This
Healthcare 2023, 11, 139
simulation utilized various nodes in this case, with node 0 serving as the source, node 1
16 of 24
serving as the HMS, node 2 serving as the FTP server, and node 3 serving as the IoT server.
Tables 3, 4, and 5 emerged from those simulations.

Healthcare 2023, 11, x 18 of 26

(a) Packet data passing

(b) Node information

Figure
Figure 11.
11. NS3
NS3 simulation
simulation from
from ULSM
ULSM to
to HMS. (a)Packet
HMS. (a) Packetdata
datapassing;
passing; (b)Node
(b) Nodeinformation.
information.
Healthcare 2023, 11, 139 (b) Node information 17 of 24

Figure 11. NS3 simulation from ULSM to HMS. (a)Packet data passing; (b)Node information.

Healthcare 2023, 11, x 19 of 26

(a) Packet data passing

(b) Node information

Figure 12.
Figure 12. NS3
NS3 simulation
simulation from
fromHMS
HMSto
toFTP
FTPand
andIoT.
IoT.(a)
(a)Packet datapassing;
Packet data passing;(b)
(b)Node infor-
Node information.
mation.

Image transfer response time was based on NS3 for performance evaluation, and the
proposed model used three metrics: throughput, packet receipt rate and received packet
size. This study calculated request response time based on packet size and bandwidth to
trace the data transfer protocol from source to destination on the IoT server.
Healthcare 2023, 11, 139 18 of 24

Table 3. ULSM to HMS (Intranet).

Data/Packet Size (MB) Bandwidth (Mbps) Response Time (s)

1 43.3442083
5 12.40370939
5 10 8.816277328
15 7.366890492
20 3.967757723
1 84.60675146
5 17.92090079
10 10 10.41828778
15 8.005866191
20 5.078943865
1 123.7062426
5 25.7373848
15 10 15.89119732
15 9.30444808
20 7.922080678

Table 4. HMS To FTP (Internet).

Data/Packet Size (MB) Bandwidth (Mbps) (S) Response Time (s)

1 41.64960943
5 10.92773353
5 10 5.634109647
15 3.674797321
20 4.611541816
1 81.17746564
5 17.22091601
10 10 10.01965412
15 8.12423431
20 6.908335693
1 122.5305769
5 25.36579175
15 10 14.6547472
15 9.84765541
20 7.490648534

Table 5. HMS To IoT (Internet).

Data/Packet Size (Byte) Bandwidth (Mbps) (S) Response Time (s)

93.2324224 1 1.252205553
89.36158621 5 1.726590514
94.16620009 10 1.217828047
94.49036105 15 0.5540504
89.44453449 20 0.738210043
Healthcare 2023, 11, x 20 of 26

Healthcare 2023, 11, 139 19 of 24

1 84.60675146
Image transfer response time was based 5 on NS3 for performance 17.92090079
evaluation, and the
10 10 10.41828778
proposed model used three metrics: throughput, packet receipt rate and received packet
size. This study calculated request response 15 time based on packet8.005866191
size and bandwidth to
trace the data transfer protocol from source 20 to destination on the IoT server.
5.078943865
The outcomes are presented in Tables 1 3–5, where the data signify
123.7062426the request and
response results depending on bandwidth 5 and packet size. Furthermore,
25.7373848 Tables 3 and 4
indicate the reaction time when packet data are sent from ULSM to HMS and from HMS to
15 10 15.89119732
FTP server, respectively. Based on the bandwidth and packet size, this study computed the
15 9.30444808
response times.
The data shown in Tables 3 and 4 20 7.922080678
are displayed in the line charts in Figure 13, which
implies that the response time and the package time increases. Additionally, when band-
Table
[Link],
HMS To FTP (Internet).
response time decreases. Again, Table 5 shows the response time from the
HMS to the IoT
Data/Packet Size server. From this outcome, we can see that response time is at a minimum
below(MB) Bandwidth (Mbps) time in Figure(S)
3 s and, in line with that, the response 13cResponse
is almost Time (s) and not
constant
dependent on bandwidth because of the minimum packet size. To decrease the response
1 41.64960943
time, this model has transferred telemetry data and an encrypted image link to the IoT
5 10.92773353
server where encrypted image links are below 100 bytes.
5 10 5.634109647
Our proposed method is compared with others of the best health monitoring methods
on the following criteria: (i) what15types of data they are using,3.674797321
(ii) the machine they are
collecting data from, (iii) if there20 is any machine-to-machine based 4.611541816
communication, (iv)
whether it supports real time monitoring,
1 and (v) is there any IoT81.17746564for displaying
dashboard
patient diagnostic information. Using 5 those criteria, the proposed method is compared
17.22091601
with other
10 existing methods in Table
10 6. 10.01965412
With the use of the internet,15we may access any LIS data8.12423431 from anywhere. In the
IoT server, we can customize the dashboard or device access by sharing patient data in
20 6.908335693
a variety of ways, including public, private, and personalized ways. Additionally, we
1 122.5305769
have computed data source links using encryption, transmitted patient information, sent
time data, and sent total data over5 time for the particular criterion. 25.36579175
The results are based
15 10 14.6547472
on how quickly the data are supplied to the IoT server, and the total number of images
transmitted each minute. When 15 we compare the criteria, we can 9.84765541
see that the amount of
data exchanged is dependent on time 20 and is secure. This data line is directly connected to
7.490648534
the ultrasonic machine.

HMS to FTP
150

100

50

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

HMS to FTP Response Time (s)

HMS to FTP Bandwidth (Mbps)
HMS to FTP Data / Packet Size (Byte)

(a)

Figure 13. Cont.

Healthcare 2023,11,
Healthcare2023, 11,139
x 20 of
21 of 24
26

ULMS To HMS (Intranet)

140
120
100
80
60
40
20
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

ULMS To HMS (Intranet) Response Time (s)

ULMS To HMS (Intranet) Bandwidth (Mbps)
ULMS To HMS (Intranet) Data / Packet Size (MB)

(b)

HMS To IoT (Internet)

100
80
60
40
20
0
1 2 3 4 5

HMS To IoT (Internet) Data / Packet Size (Byte)

HMS To IoT (Internet) Bandwidth (Mbps)
HMS To IoT (Internet) Response Time (s)

(c)
Figure 13.
Figure 13. Line
Line chart
chart diagram
diagram based
based on
on the
the response
responsetime
time(a–c).
(a–c).

[Link]-side
HMS To IoTmodule
(Internet).
is used as the secure ultrasound real-time data channel and
controls how ultrasound frame data are sent from the server to the client [20]. This data
Data/Packet Size (Byte) Bandwidth (Mbps) (S) Response Time (s)
line connects directly to the ultrasonic machine, while parameters are managed via the
93.2324224 1 1.252205553
secure real-time control route. Furthermore, the authors of one article used the LoRaWAN
protocol for data transfer, whereas we merely produced byte data 1.726590514
89.36158621 5 via URL encryption
on the IoT94.16620009
server, which resulted in a long10data response time on1.217828047
the IoT server [21].
94.49036105 15 0.5540504
Another article on sensor utilization used a smart healthcare system to collect data from the
hospital environment
89.44453449by monitoring fundamental 20 patient health indicators such as heart
0.738210043
rate and temperature, whereas we use small packet data for displaying and generating large
data with such health
Our proposed indicators
method and image
is compared sources
with others[22]. Another
of the studymonitoring
best health we encountered
meth-
gave
ods ona process for creating
the following criteria:professional, easy-to-understand
(i) what types visualizations
of data they are using, of the data
(ii) the machine they
for seamless integration into clinical practice, whereas we have successfully developed
are collecting data from, (iii) if there is any machine-to-machine based communication, a
Healthcare 2023, 11, 139 21 of 24

request-response-based data transfer and thus were able to reduce the data collision rate
to a tolerable level [23]. In another piece of research, they attempted to adapt a system in
an automated manner to follow the condition of patients when physicians are absent by
employing SHS technology, which ensures that therapy is administered to a person in the
absence of a doctor [24,25], which is quite dangerous. However, according to our system,
patients do not need to rely on such risky treatment; they can easily send their issues in a
request, and the system will store those data in a dashboard where doctors can monitor
their registered patients’ issues and provide feedback on them from anywhere and at any
time. The main hardware components are all based on the Raspberry Pi 3 Model B [26], and
it enhances overall healthcare delivery by allowing patients and clinicians to interact in a
variety of ways. Additionally, we have used only RSPI3 and large packet data as encrypted
image links over FTP for displaying data in the IoT server, which consumes less time.

Table 6. Comparison with existing methods.

Machine M to M Com- Real time IoT

Study Data Type
type munication Monitoring Dashboard
[6] Video Ultrasound No Yes No
[13] Text Sensors No No No
[15] Text Sensors Yes No No
[17] Text Sensors Yes Yes No
[22] Text Sensors Yes Yes Yes
Proposed Ultrasound,
Text, Image Yes Yes Yes
Method MRI, CT

In contrast to the other studies above, our system overcame the large data transfer
issue and displayed the image data, which may be a large amount of data. The response
was returned on time and securely. Now we will discuss the conclusions of our proposed
method and focus on whether it is the best method and if it is suitable for an IoT server.

4. Conclusions
Technology is constantly improving around the world, which is how IoT is becoming
a more practical form of technology. Every industry is becoming more advanced because
of IoT. The healthcare industry has also embraced IoT, and this paper is a straightforward
reflection of how IoT may benefit the industry. Because many people may not have the
opportunity to visit world-class healthcare experts, we created an IoT-based Ultrasound
Image Monitoring System that allows healthcare experts or doctors to easily monitor and
diagnose any patient from anywhere in the world, thanks to the benefits of cloud-based IoT
systems. Our proposed method is applied over IoT from a hospital database, so it has some
limitations e.g., for large data transfers it takes more time in real-time than the simulation,
and the response time depends on the internet bandwidth.
In this approach, both doctors and patients will be able to save time and avoid
superfluous costs such as making appointments with doctors, transportation fees, and
other costs just by using our system properly. We hope that the technology we developed,
as well as related research, will eventually bring about progress in the healthcare industry.
Further modifications and improvements to the system will improve the overall service of
the healthcare business around the world.

Author Contributions: Conceptualization—M.H.-A.-R. and methodology—O.C.; formal analy-

sis and resources—M.M.R. and M.M.H., O.C.—Chowdhury; writing—original draft preparation;
writing—review and editing, M.S.H., G.M., S.A.A., A.Y. and M.A.; supervision, M.S.H. and G.M.;
project administration, M.S.H., G.M., S.A.A. and M.A.; funding acquisition, M.S.H. All authors have
read and agreed to the published version of the manuscript.
Healthcare 2023, 11, 139 22 of 24

Funding: This work was supported by the Researchers Supporting Project number (RSP2023R32),
King Saud University, Riyadh, Saudi Arabia.
Acknowledgments: This work was supported by the Researchers Supporting Project number
(RSP2023R32), King Saud University, Riyadh, Saudi Arabia.
Conflicts of Interest: The authors declare no conflict of interest.

Appendix A

Table A1. Connection message.

ARCHITECT-Initiated Connection Test:

MSH|ˆ~\&|ARCHITECT|12345|||20091215153540+0000||NMDˆN02ˆNMD_N02|d36bb8b2-
8297-4372-be8a-87f38a86e739|P|2.5.1||||||UNICODE UTF-8SFT|ADD|8.0|ARCHITECT|”“
NST|N
The following message is an example of an Acknowledgement to ARCHITECT initiated
Connection Test:
MSH|ˆ~\&|||||20091215153541||ACKˆN02ˆACK|Ad36bb8b2-8297-4372-be8a-
87f38a86e739|P|2.5.1||||||UNICODE UTF-8
MSA|AA|d36bb8b2-8297-4372-be8a-87f38a86e739

Table A2. Order message.

Order Message
<SB>
MSH|ˆ~\&|Machine Name |Machine Model| Source | Destination| Date Time||Order Code |
Machine Model |P|2.3.1||||||UNI CODE
PID| Patient Type|Patient Name|Patient ID|Sample ID|Test Type||Patient Date of Birth|
Patient Gender<CR>
ORC|1|Requist ID|Sample ID|MachineNameˆMachine Model||Sample Time| Start
Time||||Test_Code |||||Li||<CR>
ORC|2|Requist ID|Sample ID|MachineNameˆMachine Model||Sample Time| Start
Time||||Test_Code |||||Li||<CR>
... ... .
ORC|ORDER_SERIAL|Requist_ID|Sample ID|MachineNameˆMachine Model||Sample Time|
Start Time||||Test_Code |||||Li||<CR>
<EB><CR>

Table A3. Result message.

Result Message
<SB>
MSH|ˆ~\&|Machine Name |Machine Model| Source | Destination| Date Time||Order Code |
Machine Model |P|2.3.1||||||UNI CODE
PID| Patient Type|Patient Name|Patient ID|Sample ID|Test Type||Patient Date of Birth|
Patient Gender<CR>
OBR|1|Requist ID|Sample ID|MachineNameˆMachine Model||Sample Time| Start
Time||||||Test Name || |Li||011<CR>
OBX|1|ED|IMG1|Image1|Machine ModelˆImageˆJPEGˆBase64ˆ . . . . . .
Base64_Code||||||F|||Item Result Time|011|Doctor Name|<CR>
<EB>
<CR>
Healthcare 2023, 11, 139 23 of 24

Table A4. Acknowledgement message.

The Following Message Is an Example of an Acknowledgement to ARCHITECT Initiated

Connection Test:
MSH|ˆ~\&|||||20091215153541||ACKˆN02ˆACK|Ad36bb8b2-8297-4372-be8a-
87f38a86e739|P|2.5.1||||||UNICODE UTF-8
MSA|AA|d36bb8b2-8297-4372-be8a-87f38a86e739

Appendix B

Table A5. Private key.

Private Key
MIIBVQIBADANBgkqhkiG9w0BAQEFAASCAT8wggE7AgEAAkEAnXeCBmYdiea+
DFVtFIPOEZYGdnozXJhSyMnhynYScbk1rVGyZYcWHStjRlmmra2h1Y7APmYqKljcnUgei+GzewIDAQABAkArkYZraNOhdTN+
TCCbPYDFwuHU5CjT5N166sjLcPHXfvyFWu7I3SVHw3h1jOYdsMF4pJWGrq65KWQiWpbKXgRhAiEA381fUq9LaBxevvOwRaVGK
eszA2jxXd5qfrXrB31awusCIQC0HwFLZihHl9YVaDNUbmwGtmO52eaDFVXnQq6iD/ENsQIhAKcTzRwPXb4ln93yUyBLWGwm+
HiNOdQYHWznJsT6om1tAiBB8/xoXF6xYFJ+
gioRZ2Fcz9oSSkxSgTR0OoFxS/8K8QIhAKjeIKNR3zUsPDj4JK/XVlFXD5/vV0eI8gCO4x1dLPJl

Table A6. Public key.

Public Key
MFwwDQYJKoZIhvcNAQEBBQADSwAwSAJBAJ13ggZmHYnmvgxVbRSDzhGWBnZ6M1yYUsjJ4cp2EnG5Na1RsmWHFh0rY0ZZp
q2todWOwD5mKipY3J1IHovhs3sCAwEAAQ==

Table A7. Message to be encrypted.

Message to Be Encrypted
[Link]

Table A8. Output/message to decrypt: [Encrypted Output on Base64].

Output/Message to be Decrypt: [Encrypted Output on Base64]

XHcVipE1xGw7291Kj9jJksm5xMbzQDxYhWb/IPXgyC3bjEEKzCPZr60ChsoPN2E/+HKD3i9/fIb5VmrJRUrt7w==

Table A9. Decrypted output.

Decrypted Output
s[Link]

References
1. Khan, I.I.; Ahmed, M.; Ahmad, K.U. Towards A Smart Hospital: Automated Non-Invasive Patient’s Discomfort Detection in
Ward Using Overhead Camera. In Proceedings of the 2018 9th IEEE Annual Ubiquitous Computing, Electronics & Mobile
Communication Conference (UEMCON), Columbia University, NY, USA, 8–10 November 2018; pp. 872–878.
2. Andriole, K.P.; Wolfe, J.M.; Khorasani, R.; Treves, S.T.; Getty, D.J.; Jacobson, F.L.; Steigner, M.L.; Pan, J.J.; Sitek, A.; Seltzer, S.E.
Optimizing analysis, visualization, and navigation of large image data sets: One 5000-section CT scan can ruin your whole day.
Radiology 2011, 259, 346–362. [CrossRef] [PubMed]
3. Muhammad, G.; Hossain, M.S. COVID-19 and Non-COVID-19 Classification using Multi-layers Fusion From Lung Ultrasound
Images. Inf. Fusion 2021, 72, 80–88. [CrossRef] [PubMed]
4. Krishna, K.D.; Akkala, V.; Bharath, R.; Rajalakshmi, P.; Mohammed, A.M.; Merchant, S.N.; Desai, U.B. Computer aided
abnormality detection for kidney on FPGA based IoT enabled portable ultrasound imaging system. Irbm 2016, 37, 189–197.
[CrossRef]
Healthcare 2023, 11, 139 24 of 24

5. Said, O.; Tolba, A. SEAIoT: Scalable e-health architecture based on Internet of things. Int. J. Comput. Appl. 2012, 59, 44–48.
[CrossRef]
6. Liteplo, A.S.; Noble, V.E.; Attwood, B. Real-time video transmission of ultrasound images to an iPhone. Crit. Ultrasound J. 2010, 1, 105.
[CrossRef]
7. Khairi, T.W.; Al-zubidi, A.F.; Ahmed, E.Q. Modified Multipath Routing Protocol Applied On Ns3 Dcell Network Simulation
System. Int. J. Interact. Mob. Technol. 2021, 15, 208. [CrossRef]
8. Maygua-Marcillo, L.; Urquiza-Aguiar, L.; Paredes-Paredes, M. Creating a new model in NS3 Network Simulator. Preprints
2019, 2019010106. [CrossRef]
9. Jonveaux, L. Arduino-like development kit for single-element ultrasound imaging. J. Open Hardw. 2017, 1, 3. [CrossRef]
10. Muhammad, G.; Hossain, M.S.; Kumar, N. EEG-Based Pathology Detection for Home Health Monitoring. IEEE J. Sel. Areas
Commun. 2021, 39, 603–610. [CrossRef]
11. Alshehri, F.; Muhammad, G. A Comprehensive Survey of the Internet of Things (IoT) and AI-Based Smart Healthcare. IEEE
Access 2021, 9, 3660–3678. [CrossRef]
12. Srinivasan, C.R.; Charan, G.; Babu, P.C.S. An IoT based SMART patient health monitoring system. Indones. J. Electr. Eng. Comput.
Sci. 2020, 18, 1657–1664. [CrossRef]
13. Ru, L.; Zhang, B.; Duan, J.; Ru, G.; Sharma, A.; Dhiman, G.; Gaba, G.S.; Jaha, E.S.; Masud, M. A detailed research on human health
monitoring system based on internet of things. Wirel. Commun. Mob. Comput. 2021, 2021, 5592454. [CrossRef]
14. Ali, Z.; Muhammad, G.; Alhamid, M.F. An automatic health monitoring system for patients suffering from voice complications in
smart cities. IEEE Access 2017, 5, 3900–3908. [CrossRef]
15. Anand, L.V.; Kotha, M.K.; Kannan, N.S.; Kumar, S.; Meera, M.R.; Shawl, R.Q.; Ray, A.P. Design and development of IoT based
health monitoring system for military applications. Mater. Today Proc. 2020. [CrossRef]
16. Deepa, V. Analysis on E Healthcare Monitoring System with Iot and Big Patient Data. Int. J. 2021, 5, 97–102. [CrossRef]
17. Garbhapu, V.V.; Gopalan, S. IoT based low cost single sensor node remote health monitoring system. Procedia Comput. Sci. 2017,
113, 408–415. [CrossRef]
18. Hadidi, R.; Cao, J.; Woodward, M.; Ryoo, M.S.; Kim, H. Real-time image recognition using collaborative iot devices. In Proceedings
of the 1st on Reproducible Quality-Efficient Systems Tournament on Co-designing Pareto-efficient Deep Learning, Williamsburg,
VA, USA, 24 March 2018; p. 1.
19. Available online: [Link] (accessed on 27 July 2022).
20. Tobias, R.J. Wireless communication of real-time ultrasound data and control. In Medical Imaging 2015: Ultrasonic Imaging and
Tomography. Int. Soc. Opt. Photonics 2015, 9419, 94190M.
21. Dong, M.; Sun, Z.; Yu, H.; Zhang, L.; Sui, Y.; Zhao, R. Performance Evaluation of the Transmission Quality of the Internet of
Things in Farmland Environmental Monitoring Based on ns-3. Mob. Inf. Syst. 2022, 2022, 3148348. [CrossRef]
22. Islam, M.; Rahaman, A. Development of smart healthcare monitoring system in IoT environment. SN Comput. Sci. 2020, 1, 185.
[CrossRef]
23. Hossain, M.S.; Muhammad, G. Emotion-Aware Connected Healthcare Big Data Towards 5G. IEEE Internet Things J. 2018, 5,
2399–2406. [CrossRef]
24. Masud, M.; Gaba, G.S.; Choudhary, K.; Hossain, M.S.; Alhamid, M.F.; Muhammad, G. Lightweight and Anonymity-Preserving
User Authentication Scheme for IoT-based Healthcare. IEEE Internet Things J. 2022, 9, 2649–2656. [CrossRef]
25. Gaur, L.; Bhatia, U.; Jhanjhi, N.Z.; Muhammad, G.; Masud, M. Medical Image-based Detection of COVID-19 using Deep
Convolution Neural Networks. Multimed. Syst. 2022, 1–10. [CrossRef] [PubMed]
26. Swaroop, K.N.; Chandu, K.; Gorrepotu, R.; Deb, S. A health monitoring system for vital signs using IoT. Internet Things 2019, 5,
116–129. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.