PreprintPDF Available

Smart Neonatal Intensive Care for Urban Hospitals

Authors:
Preprints and early-stage research may not have been peer reviewed yet.
Smart Neonatal Intensive Care for Urban Hospitals
Puneeth C
Information Technology,
Christ (Deemed to be University),
Bangalore, India
pnthraj37@gmail.com
Sonal Kumar
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
skbrahee@gmail.com
Naman Singhal
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
prince.naman1999@gmail.com
Samden Lepcha
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
sam.lepcha@outlook.com
Sreyan Ghosh
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
gsreyan@gmail.com
Anurag Kumar Saw
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
kanurg101@gmail.com
Aynur Unal
Digital Monozukuri,
USA
aynurunal@alumni.stanford.edu
Suraj S Jain
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
surajsjain@hotmail.com
Abhijeet Singh Batra
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
abhijeetbatra29@gmail.com
Samiksha Shukla
Computer Science and Engineering,
Christ (Deemed to be University),
Bangalore, India
samiksha.shukla@christuniversity.in
Abstract—Every year, an increasingly large number of neonatal
deaths occur in the world. Premature birth and asphyxia are
being two of the leading causes of these neonatal deaths. A well-
regulated thermal environment is critical for neonatal survival.
In the current scenario, it is difficult for the hospitals in the
urban regions to accommodate for every newborn child due to
the cost of the incubators by which the hospitals are not able to
afford many incubators, and also there are not many doctors to
attend all the neonatal cases simultaneously. The successful
delivery of neonates is hampered due to the increasing
population along with limited technology and resources. Thus, a
prototype of an incubator has been designed that is affordable,
automated, and comes along with an AI-based decision support
system to help the doctors and nurses of the hospitals so that
they get to attend to all the cases easily.
Keywords—Neonatal Incubator, Neonatal Mortality, Decision
Support System, Smart Medical Electronics, Urban Hospitals.
I. INTRODUCTION
The most vulnerable time for an infant’s survival is the first
twenty-eight days in the life of a newborn – the neonatal
period. Newborns face the highest death risk in their first
month of life. In 2017, eighteen deaths every one thousand
lives born was the average global newborn death rate.
Globally in 2017, 2.5 million children died in their first month
– approximately seven thousand neonatal deaths every day –
most of which occurred in the first week, with about a million
dying on their first day and close to a million in the next few
days.
Globally, every year almost 2.6 million babies do not survive
the first month of birth. That means an average of 7,000
neonates die every day. It has been noted that if every country
brought its newborn mortality rate to the high-income
average, or below, by 2030, 16 million newborn lives could
be saved. A study reveals that the causes of 80 percent of
neonatal deaths are preventable and treatable.
Therefore, in this paper, we will discuss our approach to solve
this problem with not only a product but also with a system
design that incorporates the various underlying processes as
a whole.
II. THERMOREGULATION OF NEONATES
There must be a balance that needs to be ensured between the
production and loss of heat by the body of the neonates. Heat
loss might affect the functions of various important organs as
well as the metabolic activity in the body of the neonates. The
loss of heat from the body of the neonates happens mostly
through the skin as the skin will be in direct contact with the
external environment. Also, if the air around the infant is
cold, it circulates within the body of the neonates as they
breathe it in and this will increase the rate of heat loss from
the bodies of the neonates.
The optimum temperature of the neonates’ body is between
36.5 °C to 37.5 °C. This optimum temperature must be
maintained in the body of the neonates in order to ensure
proper metabolism in the body. If the body temperature falls
below or exceeds the optimum temperature range, it leads to
life-threatening complications such as hypothermia and
hyperthermia that occurs when the temperature of the
neonate’s body falls below and exceeds the optimum range
respectively. These conditions are very dangerous due to the
complications that they cause in the body of the neonates.
The humidity of the surroundings of the body of the neonates
must be maintained above seventy percent in order to ensure
the proper functioning of the body. Therefore, it becomes
critical to maintain and continuously monitor an environment
that is required to maintain the body temperature as well as
ensure the comfort for the neonates.
III. PROPOSED DESIGN OF THE NEONATAL INCUBATOR
The incubator works on the principle of automating the
process of setting the thermal and humidity conditions of the
incubator by using sensors and cameras to continuously
monitor the current conditions in the incubator, and simple
actuators to adjust the conditions of the incubator in order to
best suit the baby based on the readings by the sensors and
cameras.
Temperature and humidity sensors will be used to monitor the
conditions in the incubator. High Definition and Infrared
cameras will also be used to monitor the body temperature,
heartbeat, and breathing conditions of the baby in the
incubator.
A cheap mini-computer like a Raspberry Pi will be used to
take the inputs from various sensors, the video feed from the
cameras, and send it to the cloud continuously. Cheap
computer hardware onboard the incubator and cloud
processing will be used as it would reduce the initial costs for
the hospitals to buy the incubator, so that the hospitals can
afford more incubators initially, and then pay for the cloud
service provider according to the usage. In this way, hospitals
will never run short of incubators whenever needed. Also,
cloud providers such as the Google Cloud Platform (GCP)
provide good hardware such as Tensor Processing Units
(TPUs) that are suitable to run computationally expensive
programs. The cloud also collects data from different
incubators which will be presented to the doctors and nurses
by means of an interactive smartphone/tablet/web
application.
The mini-computer will be connected to the internet all the
time and the conditions monitored in the incubator as well as
the video feed from the cameras will be sent to the cloud all
the time.
The cloud will be running the brain of the system. The brain
is the decision support system of the incubators. The brain
will receive the inputs sent by the incubator, determine the
conditions of the baby in the incubator, and send instructions
to the incubator to set the temperature and humidity
conditions of the incubator automatically to best suit the
baby, without the intervention of the doctors or the nurses.
The brain also alerts the doctors and nurses whenever
required, and also gives a sequence of instructions for the
doctors and nurses to follow in order to help them with their
decisions. This eliminates the need for continuous monitoring
of the incubator by the doctors and nurses as they can view
the details of all the incubators at one place, that is in the app
provided to them and also get notified by the app in case of
emergency. The app also lets the doctors and nurses to set
their own conditions in the incubator by overriding the
decisions made by the Artificial Intelligence, depending on
the readings as Artificial Intelligence models in the brain of
the system cannot be relied upon all the time.
The mini-computer also receives instructions from the cloud
that is either sent by the brain, by doctors or nurses from their
app and controls the computer fans and incandescent light
bulbs to regulate and maintain the temperature inside the
incubator. A water heater will also be controlled by the mini-
computer to produce water vapor to maintain the humidity
inside the incubator. A thermal bed would be used for the
baby in the incubator as it would help in maintaining the
baby’s body temperature. The sensors will sense the
conditions in the incubator, send it to the cloud, the brain
running in the cloud or the medical professional will decide
what is the best temperature and humidity conditions that are
to be set, the cloud sends back the conditions to the mini-
computer in the incubator that controls the heating, cooling
and humidity control systems to maintain the temperature and
humidity conditions in the incubator.
The body of the incubator will be built using biodegradable
plastic. Biodegradable plastic is made out of fermented starch
from plants such as corn, cassava, sugarcane or sugar beets.
Unlike the other plastics made from petroleum,
biodegradable plastics can be easily recycled, and also they
have the same properties as other plastics such as heat
resistance and durability so that the conditions inside the
incubator can be maintained effectively, and also the
incubator’s body can be made without any sharp edges.
The incubator will also be provided with a battery to back the
electricity. This can be useful if the incubator is to be moved
between rooms in the hospitals without the operations being
interrupted.
Fig. 1: Outside view of the incubator.
Fig. 2: Right side diagonal view of the incubator and the parts inside.
Fig. 3: Left side view of the incubator and the parts inside.
A. RASPBERRY PI 3 MODEL B+
The raspberry pi 3 model b+ would be used as a
processor in the incubator. This can be replaced by any
other cheap processor that can run a Linux distribution
operating system on it. The raspberry pi used here will
be running Raspbian OS.
The processor will be used to interface with the sensors,
fans, cameras, and a relay module for interfacing with
the water heater and the incandescent bulb.
The computer will continuously fetch the readings from
the sensors and the footage from the camera for
determining the conditions of the baby as well as the
conditions inside the incubator and send it to the cloud
over the internet. The cloud will be running the brain of
the system that will determine the conditions that needs
to be maintained inside the incubator. Doctors and nurses
will also have access to a mobile/web application that
shows them the current conditions and previous
conditions of the incubator as well as the previous
actions taken by the brain of the system. The doctors and
nurses can override the actions taken by the brain of the
system and set their own conditions. When the doctors
or nurses set their own conditions, the conditions will be
sent to the cloud, the back-end program on the cloud will
stop the brain and follow the conditions set by the
doctors or the nurses. The cloud will send the conditions
set either by the brain, or by the doctors or nurses to the
Raspberry Pi.
The Raspberry Pi reads the conditions that is to be
maintained and control the heating, cooling and water
heater to maintain the temperature and humidity
conditions set by the brain of the system.
B. COOLING SYSTEM
The cooling system would be used to decrease the
temperature inside the incubator when the baby’s body
temperature is rising up so as to maintain the baby’s body
temperature between 36.5°C and 37.5°C.
The cooling system would use two 12v DC fans
controlled by the Raspberry Pi. Whenever there is a rise
in body temperature of the baby, it will turn the cooling
fans on. The cooling fans would then take the hot air out
of the incubator, which would decrease the temperature
inside the incubator and cool the baby's body down so
that the optimum temperature is maintained.
C. ACRYLIC COVERING
The covering of the incubator will be made of a thin and
light acrylic sheet. It is easy to open and close.
Acrylic is highly durable and a good heat insulator. This
allows the incubator resistant to damage when used in
ambulances and also resistant to rough usage. The
temperature inside the incubator would not be affected
easily by the outside temperature due to the acrylic
covering, which makes it easy to maintain the
temperature and the humidity inside the incubator
efficiently.
D. LOAD SENSOR
The load sensor would be placed on the bottom of the
inner surface of the incubator. It would be connected to
the computer and gives the bodyweight of the baby to the
computer. The body weight can be used as a feature for
the brain (composed of various ANNs) running on the
cloud to determine the baby’s medical conditions.
E. CAMERA
The incubator would use two cameras. One of the
cameras will be a cheap webcam with a microphone. The
input from this camera will be fed as an input to a
Convolutional Neural Network (CNN) running in the
brain on the cloud to determine the skin color and other
behavioral patterns of the baby. This can be used to
detect jaundice and other conditions of the baby. The cry
of the baby can be monitored by the microphone and can
be fed into another ANN to determine other behavioral
patterns and conditions of the baby.
The other camera will be an infrared camera. This can be
used to measure body temperature, heartbeat, and
breathing patterns of the baby. This camera’s input will
be fed into different Convolutional Neural Networks
(CNN) running in the computer to determine the body
temperature, heartbeat and breathing patterns of the
baby, and the conditions inside the incubator will be set
accordingly to best suit the baby.
F. HEATING SYSTEM
The heating system comprises a computer fan and an
incandescent bulb. Whenever a fall in the body
temperature of the baby is detected by the brain running
on the cloud, the brain sends a message to the Raspberry
Pi to activate the heating system to raise the temperature
inside the incubator which would, in turn, increase the
body temperature of the baby in order to maintain it in
the optimum range.
The heating system is activated automatically by the
processor by turning on the incandescent bulb to heat the
air and the computer fan to push the hot air into the
incubator to raise the temperature in the incubator.
G. INCANDESCENT BULB
A 100W incandescent bulb will be used in the incubator
which when turned on, heats the surrounding air. The
automatic turning on and off of the incandescent bulb
would be controlled by the computer by means of a relay
module.
H. WATER STORAGE
The water storage space is a part of the humidity control
system in the incubator. It can store water that can be
heated to produce vapor. The vapor will further be
directed into the incubator in order to maintain the
humidity level in the incubator above 70 percent.
I. WATER HEATER
The water heater is a part of the humidity control system
in the incubator. It can be turned on and off by the
Raspberry Pi by means of a relay module.
The Raspberry Pi turns the water heater on automatically
when it detects a fall in the humidity level below 70
percent. The water heater will use two coils, which when
heated by electricity, heats the water in the water storage
to produce water vapor.
J. WATER INLET
The water inlet to the incubator is a part of the humidity
control system. It allows the water storage to be refilled
easily by using a pipe and a tap. A funnel can also be
attached to it in order to fill water into it from other
sources.
K. LCD TOUCH SCREEN
The LCD screen will be displaying the web version of
the app built for the doctors and the nurses. The app
displays the current conditions of the incubator, the
decisions taken by ‘The Brain’ of the system, and will
also give an option for the doctors and the nurses to
override the decisions taken by the Brain of the system
and set their own temperature and humidity conditions.
L. THERMAL SHEET
Thermal sheets help in maintaining the body temperature
of the infant in the incubator. Thermal sheets do not lose
heat easily as compared to the other sheets. This would
make the task of maintaining the conditions in the
incubator easier, and will also be more comfortable for
the baby.
IV. THE BRAIN AND THE CLOUD
Fig. 4: Overview of the role of the back end in the incubator system
The Incubator sends data continuously to the back end hosted
on the Google Cloud Platform. The data will be stored in the
database hosted on the cloud. The data contains information
such as the sensor readings of the conditions inside the
incubator, and the conditions of the baby as determined by
the cameras. Google Cloud Platform will be used as the cloud
service provider for this system as it offers hardware
optimized for running various Artificial Neural Networks
(ANNs) running on the cloud. Artificial Neural Networks are
an important part of the system, and they are also
computationally expensive to run. Usage of poor hardware
for running the Artificial Neural Networks would result in the
delay of execution which would in-turn delay the setting of
the conditions in the incubator. Google Cloud Platform
provides special hardware such as Tensor Processing Unit
(TPUs) and high-end Graphics Processing Unit (GPUs) that
is optimized to run Neural Networks. In this way, there will
not be any delay in computation and the conditions in the
incubator will be monitored and set without any delay.
The cloud will be running a program called ‘The Brain’ of
the system. The Brain consists of various Artificial Neural
Networks. These Neural Networks make use of the data that
is sent by the incubators to the cloud and produces results that
can be used to perform various actions such as controlling the
conditions in the incubator, alerting the doctors and nurses in
case of an emergency. Some of the Artificial Neural
Networks in the brain act as a decision support system by
recommending instructions that can be followed by the
doctors and the nurses during an emergency situation.
The instructions from the Brain can be viewed by the doctors
and the nurses on the mobile/web app provided to them. The
doctors and the nurses can also view the conditions of the
incubator recorded in the database hosted on the cloud
through their mobile/web app.
The doctors and nurses can also set their own conditions
based on the conditions of the baby in the incubator. They can
set the conditions easily through the mobile/web app based
on the privilege level provided to them by just entering the
value of the temperature and the humidity values on the app.
When the doctors or the nurses set their own conditions on
the mobile app, the app sets an override condition for the
incubator on the cloud. When the override condition is set,
the cloud will stop the brain temporarily, till the doctors or
the nurses turn off the override condition on the mobile/web
app. The cloud will send the conditions that the doctors or the
nurses have specified on their app to the Raspberry Pi on the
incubator.
V. MOBILE APPLICATION
The doctors and the nurses will be given a mobile app. This
app will be available for IOS as well as Android device. The
app will display the status of all the incubators in the hospital
at one place so that the doctors and nurses do not have to
move around the hospitals all the time.
The doctors and the nurses will have to login by the
credentials provided to them by the hospital in order to view
the status of the active incubators and set their own
conditions. This is done for security purposes so that only the
doctors and the nurses can set the conditions in the incubator
and not anyone else.
In some cases, the hospital can also opt for an application that
gives guest access to guest doctors and nurses. The guest
access can be assigned in a way that the guest doctors or
nurses can view or manipulate the conditions of only one or
more incubators assigned to them by the hospital. The guest
access can be set for a few hours or days by the hospital.
A web-based dashboard will be provided to the hospital
administrators where they can view the profiles of all the
doctors and nurses of the hospital, and also the guest doctors
and nurses.
Fig. 5: The login screen
After logging in, the doctors and the nurses will see the screen
that displays the incubators in which a critical condition has
been determined by the Brain in the cloud, and it also displays
the rooms in which the incubators are kept. The top bar of this
screen gives the count of the number of active incubators and
also the number of critical cases.
The doctors or the nurses can either click on the incubators in
the critical condition and view the conditions of that
incubator, or they can click on a room and view all the
incubators in the room. The incubator and the room widgets
on the screen will display the live video feed captured from
the cameras in the incubator and the room in which the
incubators are.
When the doctors or the nurses click on a room that they want
to view, they will be redirected to the screen that displays the
widgets of all the incubators within that room.
When the doctors or the nurses click on the widget of any
incubator, they will be redirected to the screen that displays
all the conditions of the incubator as well as the baby inside
the incubator, and also the doctors and the nurses assigned to
work on that incubator. They can also set their own
conditions for the incubator by clicking on the temperature or
the humidity value under the incubator conditions, and then
clicking on the set button. This will automatically set the
override condition on the cloud. The doctor or the nurse
viewing this screen can scroll down, click ‘Attend this case’,
and they will be assigned to that case. This makes it easier for
the doctors and the hospital to view and manage all the cases
at one place.
Fig. 6: Screen 2 displays the widgets of the incubators of critical cases
and also the widgets of the rooms in which the incubators are.
Fig 7. The app screen that displays the widgets of all the incubators in
a room.
Fig 8. The app screen that displays the conditions inside the incubator
as well as the conditions of the baby.
VI. COST ANALYSIS OF THE INCUBATOR
The incubator we have designed uses cheap and readily
available components used for Do It Yourself (DIY) projects.
This allows the incubator to be fixed easily and in a faster
way by just replacing the necessary components in case of
breakage.
The software part of the incubator will be made using open-
source libraries, so there will not be any extra licensing costs
involved for running the software. The charges for the Google
Cloud Platform vary by the region where the instance is used.
A cloud instance is used so that expensive hardware need not
be used onboard the incubator which would reduce the initial
cost of the incubator so that the hospitals can afford more
incubators easily and never run out of incubators in case of
emergency.
The hospitals can pay the cloud charges only when the
incubators are used according to the demand, and also bill the
cloud charges of the incubator to the parents of the baby. This
would also be helpful in situations where insurance can be
claimed and also in countries where the medical charges are
taken care of by the government.
Item Cost Quantity Total
Raspberry Pi 3 B+ $40.00 1 $40.00
Acrylic sheet $15.00 1 $15.00
Biodegradable plastic
body $20.00 1 $10.00
Humidity, Temperature
Sensor -DHT 11 $2.00 4 $8.00
IR Camera AMG8833 $40.00 1 $40.00
USB Camera -
Logitech
c310 webcam $27.00 1 $27.00
Relay Module -
4
Channel $3.00 1 $3.00
100W Incandescent bulb
$1.50 2 $3.00
12V DC fan $3.00 4 $12.00
Bed (Thermal bed) $22.00 1 $22.00
Water heater $1.50 2 $3.00
Backup Battery $50.00 1 $50.00
500W 12v DC to 220V
AC inverter $15.00 1 $15.00
Battery charger $5.50 1 $5.50
Connections $8.00 1 $8.00
3.5' LCD touch display $26.00 1 $26.00
USB Microphone $4.50 1 $4.50
Weighing Sensor $1.50 1 $1.50
Total Cost $303.50
Table 1: A table showing various components used to build the
incubator along with the quantities and their cost in USD in India.
VII. COMPARISON WITH PREVAILING
INCUBATOR
The incubator proposed has major advantages over the other
incubator available in the market. It is significantly cost-
effective and intelligent unlike other incubators available in
the market as it is driven by AI. Acrylic sheets make the
whole incubator build stronger, lighter and easier to clean and
handle. Maintenance costs are lesser when compared to
general hospital-grade incubators. As the incubators are
designed in a simple and robust way, they are easy to carry
around in the hospitals and are even suitable to carry in a
moving vehicle like an ambulance, provided that there is an
internet connection along the ambulance route. The need for
special attention could be decreased as most of the calls are
made by the AI system built into the incubator unlike other
incubators available in hospitals today.
VIII. CONCLUSION AND FUTURE SCOPE
This neonatal incubator could provide a proper and conducive
environment for the newborn. Primary care can be provided
within a short period during the critical condition, thus
reducing the mortality case among neonatal infants. This will
also be monitored by an AI-based system to substitute the
constant supervision of doctors and nurses and assist them in
decision making. Hence, the cost and resources for the
operation will be reduced.
The proposed system has few limitations which could
possibly be overcome by continuous technology
advancement in the future. In the recent development stage,
as there could be some critical situations that are dangerous
for the life of the infant, AI may not be the best option that
people can trust upon. Although the incubator is powered by
several sophisticated batteries and power backups, in the
worst case there could be few cases of the incubator running
low on the power especially in places where finding a power
supply/backup is difficult.
IX. ACKNOWLEDGMENT
This project was started as a research initiative at the Faculty
of Engineering, CHRIST (Deemed to the University),
Bengaluru, India. We thank all the members concerned for
the same.
REFERENCES
[1] Puneeth C, Samden Lepcha, Suraj S Jain, Sonal Kumar, Abhijeet
Singh Batra, Anurag Kumar Saw, Naman Singhal, Samiksha Shukla,
Aynur Unal on “Smart Portable Neonatal Intensive Care for Rural
Regions”, unpublished.
[2] “Neonatal mortality - UNICEF DATA,” UNICEF DATA.
[Online]. Available: https://data.unicef.org/topic/child-
survival/neonatal-mortality/
[3] Jin Fei and Ioannis Pavlidis on “Analysis of Breathing Air Flow
Patterns in Thermal Imaging”, Conf Proc IEEE Eng Med Biol Soc.,
IEEE, 2006.
[4] Yoonkyoung Kim, Yosep Park, Jinman Kim, Eui Chul Lee* on
“Remote Heart R ate Monitoring Method Using Infrared Thermal
Camera”, International Journal of Engineering Research and
Technology, 2018.
[5] Robert Frischer, Marek Penhaker, Ondrej Krejcar, Marian
Kacerovsky and Ali Selamat on “Precise Temperature Measurement
for Increasing the Survival of Newborn Babies in Incubator
Environments”, Sensors (Basel), 2014
[6] Fazle Elahi and Zinat Ara Nisha on “Low Cost Neonatal
Incubator With Smart Control System”, 8th International Conference
on Software, Knowledge, Information Management & Applications,
2014.
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
A neonatal incubator has been designed, modeled and developed which is incorporated with embedded temperature and humidity control system. The chamber has two compartments; larger and smaller. Smaller compartment consists of temperature and humidity control unit and larger compartment consists of a mattress where the baby is kept. The control system components include relative sensors, fans, bulbs, heater and Arduino Uno microcontroller. For implementation, a software program has been developed in C which comprises a code editor. A flowchart is provided to illustrate the logical expression of the program. The salient features of the model are – it is entirely microcontroller based, can be locally developed at a low cost and is a sophisticated version of conventional incubator systems. The system can control the temperature at the set level of 36 o C to 37 o C and relative humidity at the set level of 70% to 75%. At the same time the incubator is free of health hazard. It can be used commercially in the hospitals as too at home.
Article
Full-text available
Precise temperature measurement is essential in a wide range of applications in the medical environment, however the regarding the problem of temperature measurement inside a simple incubator, neither a simple nor a low cost solution have been proposed yet. Given that standard temperature sensors don't satisfy the necessary expectations, the problem is not measuring temperature, but rather achieving the desired sensitivity. In response, this paper introduces a novel hardware design as well as the implementation that increases measurement sensitivity in defined temperature intervals at low cost.
Article
The heart rate and heart rate variability are types of biometric information that are used in healthcare and wellness applications. Although attached-sensor-based photoplethysmography methods are widely used, they are not suitable for athletes, patients, and babies, who have difficulty wearing sensors. Generally, the human skin exhibits minute temperature changes due to blood flow. These changes are challenging to visually grasp, even with a thermal camera. In this paper, a thermal camera and some image processing methods are used to amplify the minute temperature changes due to facial-skin blood flow, and the heart rate is calculated using the amplified temperature signal. Results showed that the proposed method estimated heart rate of about 95% accuracy in cases of accurately detected blood vessel.
Article
We introduce a novel methodology to characterize breathing patterns based on thermal infrared imaging. We have retrofitted a Mid-Wave Infra-Red (MWIR) imaging system with a narrow band-pass filter in the CO(2) absorption band (4130 - 4427 nm). We use this system to record the radiation information from within the breathing flow region. Based on this information we compute the mean dynamic thermal signal of breath. The breath signal is quasi-periodic due to the interleaving of high and low intensities corresponding to expirations and inspirations respectively. We sample the signal at a constant rate and then filter the high frequency noise due to tracking instability. We detect the breathing cycles through zero cross thresholding, which is insensitive to noise around the zero line. We normalize the breathing cycles and align them at the transition point from inhalation to exhalation. Then, we compute the mean breathing cycle. We use the first eight (8) harmonic components of the mean cycle to characterize the breathing pattern. The harmonic analysis highlights the intra-individual similarity of breathing patterns. Our method opens the way for desktop, unobtrusive monitoring of human respiration and may find widespread applications in clinical studies of chronic ailments. It also brings up the intriguing possibility of using breathing patterns as a novel biometric.
Aynur Unal on "Smart Portable Neonatal Intensive Care for Rural Regions
  • C Puneeth
  • Samden Lepcha
  • S Suraj
  • Sonal Jain
  • Abhijeet Kumar
  • Anurag Kumar Singh Batra
  • Naman Saw
  • Samiksha Singhal
  • Shukla
Puneeth C, Samden Lepcha, Suraj S Jain, Sonal Kumar, Abhijeet Singh Batra, Anurag Kumar Saw, Naman Singhal, Samiksha Shukla, Aynur Unal on "Smart Portable Neonatal Intensive Care for Rural Regions", unpublished.