VOICE RECOGNITION BASED MEDI ASSISTANT

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 10 Issue: 04 | Apr 2023 www.irjet.net p-ISSN: 2395-0072
© 2023, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 1526
VOICE RECOGNITION BASED MEDI ASSISTANT
Prof. Smitha1, Mr. Abin V V2, Mr. Deviprasad3, Mr. Srijan R Mahale4, Mr. Vishal5,
Ms. Shreya K6
1Asst. Professor, Dept. of CSE, YIT Moodbidri, Mangalore, Karnataka, India
2,3,4,5,6 B.E Student, Dept. of CSE, YIT Moodbidri, Mangalore, Karnataka, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The "VOICE RECOGNITION BASED MEDI
ASSISTANT" promises to improve healthcare through a
paradigm for hands-free and eye-free interaction. However,
little is known about the problems encountered during
development of system especiallyfortheelderlyhealth-related
work. To solve this let's look at maintenance delivery method
and QoL improvement for the elderly. finally, identified key
design challenges and opportunities can occur when
integrating voice-based interactions virtual assistant (IVA)in
the life of the elderly. our findings will help practitioners
conducting research and research. development of intelligent
devices utilizing his full-fledged IVA provide better care and
support a better quality of life because of the
aging population.
Key Words: Interactive Virtual Assistant(IVA), Medi assistant,
Chatbot,
1.INTRODUCTION
The aging of the population was seen as a global problem in
the 21st century approach for improving healthcare and life
routine management resulting in an improvement in quality
of life QoL however obtaining and utilising such
breakthrough technologies is frequently difficult for ageing
folks who are frequently left behind while numerous
systems such as Electronic Health Record (HER) and Patient
Portals (PPS) have been widely used to improve health data
management and patient provider communications it is
difficult for ageing populations with multiple comorbidities
to adopt learn and interact with such toolswhichcanonlybe
accessed through graphical user interfaces guis on desktops
and mobile devices. the concept of this simple but efficient
project emerged out of the need to overcome this fear of the
driver and other road users if they are properly
implemented overall this simple device could help reduce
the 40-0 at some point in the future.
Intelligent Virtual Assistants (IVAs) based on voice
allow users to naturally engage with digital systems while
remaining hands-free and eye-free. As a result, they are the
next game changer for future healthcare, particularlyamong
the elderly. With the rising acceptanceofIVAs,researchhave
examined how older persons use existing IVA features in
smart speakers to enhance their daily routines, as well as
potential impediments preventing their adoption.
While these findings mostlyfocusedonolderadults'
experiences with existing IVA features on a single type of
smart-home device, practical demands, challenges, and
design strategies for incorporating these devices into older
persons' daily lives remain unexplored. Furthermore, while
the quality of healthcare delivery and QoL enhancement are
controlled by both care providers and patients, previous
research solely focused on the patient experience, leaving a
huge gap between provider expectations and quality of care
delivery.
1.1 Objective Of Research
The goal of research for a voice and machine learning-based
medical assistant is to create a reliable system that can help
medical professionals diagnose and treat patients, give
patients personalizedhealthadvice,automateadministrative
tasks, enhancemedicaltranscriptionanddocumentation,and
improve the patient experience overall.
2. LITERATURE REVIEW
[1]Angel-Echo: A Personalized Health Care Application.
Mengxuan Ma, Karen Ai, and Jordan Hubbard are the
authors :
Technology iscontinuallybreakingnewgroundinhealthcare
by providing new avenues for medical personnel to care for
their patients. Health data can now be collected remotely
without limiting patients' independence, thanks to the
development of wearable sensors and upgraded types of
wireless communication such as Bluetooth lowenergy(BLE)
connectivity. Interactive speech interfaces, such as the
Amazon Echo, make it simple for those with less
technological skills to access a range of data byutilisingvoice
requests, making communicating with technology easier.
Wearable sensors, such as the Angel Sensor, in conjunction
with voice interactivedevices,suchastheAmazonEcho,have
enabled thecreationofapplicationsthatallowforsimpleuser
involvement. We propose a smart application that monitors
healthstatus by integratingtheAngelsensor'sdatacollection
capabilities and the Amazon Echo's voice interface
capabilities. We also give Amazon Echovoicerecognitiontest
results for various populations.

[2]Implementation of interactive healthcare advisor
model using chatbot and visualization. Tae-Ho Hwang,
JuHui Lee, Se-Min Hyun, and KangYoon Lee are the
authors:
Using the influx of various information has major affects on
human lives in the fourth industrial revolution era. The
application of artificial intelligence data in the medical
industry, in particular, has the potential to influence and
affect society. The components required for establishing the
interactive healthcare advisor model (IHAM) and chatbot-
based IHAM are described in this paper. The biological
information of the target users used in the study, such as
body temperature, oxygen saturation (SpO2), pulse,
electrocardiogram (ECG), and so on, was measured and
analysed using biological sensors based on the oneM2M
platform, as well as an interactive chatbot to analyse
everyday biophysical conditions. Furthermore, the
accumulated biological information in the chatbot and
biological sensors are sent to users via the chatbot, and the
chatbot also provides medical advice to boost the user's
overall health.
[3]Chatbot for Healthcare System Using Artificial
Intelligence. Lekha Athota, Vinod Kumar Shukla, Nitin
Pandey, and Ajay Rana are the authors:
Health care is critical to a happy life. However, it is quite
difficult to get a doctor'sconsultation to every health-related
condition. The idea is to use Artificial Intelligence to build a
medicalchatbot thatcan diagnosediseasesandprovidebasic
information about them before contacting a doctor. This will
help to minimise healthcare expenditures while also
improving access to medical knowledge via a medical
chatbot. Chatbots are computer programmes that engage
with users through natural language. The chatbot saves the
data in the database in order to recognise the sentence
keywords, make a query decision, and respond to the
question. The n-gram, TFIDF, and cosine similarity are used
to calculate ranking and sentence similarity. Each sentence
from the given input sentence will be scored, and more
similar sentences will be found for the query. The expert
programme, a third party, tackles the question supplied to
the bot that is not understood or is not in the database.
[4]IntelliDoctor – AI based Medical Assistant. Dr. Meera
Gandhi, Vishal Kumar Singh, and Vivek Kumar are the
authors:
IntelliDoctor is a personal medical assistant powered by
Artificial Intelligence (AI). This interactive programme
analysessymptomstodiagnose,anticipatemedicaldisorders,
and offers remedies and ideas dependingontheuser'sinputs
in an effort to deliver smart healthcare and make it more
accessible. Furthermore, the software tracks users' health
activities such as step counts, sleep tracking, heart rate
sensing, and other information, and provides users' periodic
health reports. It takes into account numerous exercise
activities tracked as well as other criteria such as their age,
gender, location, previous medical data,and calorieintake to
provide a more accurate analysis.accurate analysis. It
provides accurate comprehensive diagnosis and also
functions as a pre-screening instrument for doctors.
[5]verview of the Speech Recognition Technology.
Jianliang Meng, Junwei Zhang, and HaoquanZhaoarethe
authors:
Speech recognition is a cross-disciplinary field that uses the
voice as the research object. Speech recognition enables the
machine to convert a speech signal intotextorcommandsvia
an identification and understanding process, as well as to
perform natural voice communication.
2. PROPOSED WORK
Fig -1: System Flowchart
Conversational virtual assistants, or voice assistants,
automate user interactions. Artificial intelligence is used to
fuel chatbots, which uses machine learning to comprehend
natural language. The paper's primary goal is to assist
readers with basic health information. When a person
initially accesses the website, they must register before they
can ask the bot questions. If the answer is not found in the
database, the system uses anexpert system to respondtothe
requests. Domain experts are also required to register by

providing certain details. The chatbot's data is saved in the
database as pattern-template data. The database queries in
this case are handled by a NoSQL database.
Fig -2: System Architecture
Fig- 2 Shows System architecture. The inquiry is entered by
the client into the UI as speech. The user interface receives
the user's inquiry and then delivers it to the chatbot
programme. The pre-processing stages for literary
experiences in the chatbot application include tokenization,
in which the words are tokenized, the stop words are then
eliminated, and feature extraction is based on N-gram, TF-
IDF, and cosine likeness. The knowledge database stores the
answers to questions so that they can be recovered and
retrieved.
Tokenization: The word-by-word separationofsentencesor
words for easier processing. Every time it encounters one of
the rundowns of the selected character, it divides text into
words. Sentences are broken up into individual words, and
all punctuation is removed. This suggests what comes next.
Stop words removal:Stop words are eliminated from
sentences in order to extract significant keywords. It is
mostly used to eliminate extraneous elements from
sentences, such as words that occur far too frequently.
Additionally, it is utilised to remove terms like an, a, and the
that are unnecessary or have ambiguous meanings. This
action is taken to lessen computational complexity or
processing time.
N-gram TFIDF-based feature extraction: The method of
feature extraction, which ranks the qualities in accordance
with the document, is one of characteristic diminution. This
phase improves the document's efficiency and suitability. It
is employed to extract the list of keywords and their
frequency within the text.
TF-IDF: The weight of each phrase in the sentence is
determined using phrase Frequency and Inverse Document
Frequency. To determine how frequently a word or phrase
appears in a sentence, use the term frequency.
N-gram: The goal of N-gram is to expand N-gram models
through the use of variable length arrangements.Agrouping
of words, a word class, a grammatical feature, or any other
succession of items that the modeller believes to have
important language structure data might be considered a
sequence. N-grams are employed in this system to extract
the pertinent keywords from the database, compress the
text, or decrease the amount of data in the document.
Sentence similarity:Todeterminehowsimilartwosentences
are, cosine similarity is utilised. The number of query
weights directly relates to how similar the query and the
document are. Since the word frequency cannot be negative,
the similarity calculation result for the two papers falls
between 0 and 1.
Find the matching phrase: The user interface retrieves and
displays the answers to the query that were discovered
through the aforementioned process.
Results and Analysis: The application uses a question-and-
answer protocol, and it consists of a login page where users
must provide their informationtoregisterfortheapplication
if they are new users, a page that displays similar answers to
the user's query if one is already in the database, and a page
where experts respond to questions directly from users. To
speed up query execution, the application leverages bigram
and trigram in addition to n-gram text compression. To
communicate the responses to the users, N-gram, TF-IDF,
and cosine similarity were used.
Web technology in use: React is a UI development library
based on JavaScript. It is controlled by Facebook and an
open-source development community. React is a popular
library in web development even though it isn't a language.
The library made its debut in May 2013 and is currently one
of the frontend libraries for web development that is most
frequently used. The application will use Express.js for
server side development together with MongoDB as its
primary database.
3. RESULTS
Systems that engage with patients, respond to their
questions about medicine, and offerthemmedical adviceare
known as voice recognition-based medical assistants. These
technologies are intended to be more effective than more
conventional forms of communication, such phone calls or
emails, and they can give patients advice that is more
individualised and precise. Utilising machine learning
algorithms to analyse medical data andfindpatternsthatcan
be used to forecast a patient's likelihood of contracting a
specific disease is known as machine learning-baseddisease
prediction. Large datasets of medical records, genetic
information, and other pertinent data can be used to train
these algorithms to increase their accuracy and
dependability.

Fig -3: Voice Chatbot
Fig -3 shows Voice chatbot which is created with javascript,
html and CSS as templetes and runs on the Django server.
The chatbot uses web speech recognition API to provide
chatbot functionality.
Fig -4: Results After Diagnosing the Symptoms from User
Fig-4 shows that the chatbot will accept user input and
deliver it to the backend ML algorithm, which will forecast
the results. When the results are computed, the chatbot will
pronounce the output and display the results along with the
required data.
Fig -5 Menu for Doctor-Patient Consultations
Fig -5 shows the doctor and patient being able to talk about
the disease, and the doctor having complete information
about the disease projected.
Fig -6 Doctor, User and Admin Login Menus
Fig -6 shows, a user, administrator, or doctor can connect
into the system using this menu, where he or she will beable
to re-login automatically during the followingvisit.Fromthe
perspective of the logged in user, the API requests will be
secure.
5. ADVANTAGES
There are several advantages to using physician assistants
based on speech recognition and disease prediction using
machine learning in health care. Some of these
advantages include:
Improved accuracy and speed of diagnoses: Machine
learning algorithms are able to analyze large amounts of
medical data to identify trends and predict disease risk,
leading to earlier accuracy in dignoses.
Enhanced patient experience: Voice-activated medical
assistants provide patients with instant access to medical
advice and information, enabling them to manage their
health more effectively.
Increased efficiency: Routine chores canbeautomatedusing
voice recognition-based medical assistants, saving medical
practitioners time and allowing them to focus on more
sophisticated patient care.
Personalized medicine: Individual patient data can be
analysed by machine learning algorithms to produce
personalised treatment plans based on characteristics such
as heredity and lifestyle, resulting in more
successful therapies.
Reduced healthcare costs: Voice recognition-based medical
assistants and diseasepredictionusingmachinelearning can
help cut healthcare expenses by enabling earlier disease
detection and more effective treatments.

Improved healthcare outcomes:Voice recognition-based
medical assistants and illness prediction using machine
learning can enhance overall healthcare outcomes by
offering patients with more efficient, accurate, and
personalised medical treatment.
Improved healthcare outcomes:Voice recognition-based
medical assistants and illness prediction based on machine
learning can enhance overall healthcare outcomes by
offering patients with more efficient, accurate, and
personalised medical treatment.
Overall, voice recognition-based medical assistants and
illness prediction based on machine learning have the
potential to transform healthcare by offering more efficient,
accurate, and personalised medical treatment while
lowering healthcare costs and increasing patient outcomes.
6 DISADVANTAGES
Along with the benefits, there are some possible drawbacks
to employing voice recognition-based medical assistantsand
machine learning to forecast disease in healthcare. Here are
some of the major drawbacks:
Privacy and security concerns: Patient data collection and
storage may pose privacy and securityconcerns,especiallyif
the data is not properly secured or comes into
the wrong hands.
Algorithmic bias: ML algorithms can be biased if thetraining
data is biased, potentially leading to inaccurate predictions
or diagnoses and perpetuating healthcare disparities.
Limited access: Some patients may not have access to the
technology required to employ voice recognition-based
medical assistants, thus leaving them behind.
Technical difficulties: Technical issues, such as voice
recognition mistakes or software faults, could result in
inaccurate diagnosis or recommendations.
Legal and ethical issues: The employment of voice
recognition-based medical assistants and machinelearning-
based disease prediction poses legal and ethical concerns,
such as accountability and obligation in the event of an
inaccurate diagnosis or advise.
Dependency on technology: The increased reliance on
technology may result in less human interaction and
empathy, thus compromising patient satisfaction and trust.
Overall, while voice recognition-based medical assistants
and illness prediction using machine learning have the
potential to transform healthcare, it is critical to address
these possible drawbacks to ensure that they are utilised
responsibly and ethically.
7. CONCLUSIONS
Finally, by giving patients more effective,individualised,and
precise medical advice and diagnoses, voice recognition-
based medical assistants and disease prediction using
machine learning have the potential to revolutionise
healthcare. These technologies do, however, also giverise to
privacy and data security issues, as well as the risk of
algorithmic discrimination and bias. Strong privacy and
security protocols, unbiased and trustworthy disease
prediction algorithms, and the incorporation of ethical
considerations into the development and use of these
systems are all necessary for addressing these concerns in
order to ensure the success of these technologies.
Overall, the integration of these technologies offers
hope for the future of healthcare, but it is crucial toapproach
their development and deployment with care and prudence
to guarantee that they are successful.
REFERENCES
[1] Ma, M., Skubic, M., Ai, K., & Hubbard, J. (2017, July).
Angel-echo: a personalized health care application. In
2017 IEEE/ACM International ConferenceonConnected
Health: Applications, Systems and Engineering
Technologies (CHASE) (pp. 258-259). IEEE.
[2] Hwang, T. H., Lee, J., Hyun, S. M., & Lee, K. (2020,
October). Implementation of interactive healthcare
advisor model using chatbot and visualization. In 2020
International Conference on Information and
Communication Technology Convergence (ICTC) (pp.
452-455). IEEE.
[3] Athota, L., Shukla, V. K., Pandey, N., & Rana, A. (2020,
June). Chatbot for healthcare system using artificial
intelligence. In 2020 8th International conference on
reliability, infocom technologies and optimization
(trends and future directions)(ICRITO) (pp. 619-622).
IEEE.
[4] Gandhi, M., Singh, V. K., & Kumar, V. (2019, March).
Intellidoctor-ai based medical assistant. In 2019 Fifth
International Conference on Science Technology
Engineering and Mathematics (ICONSTEM) (Vol. 1, pp.
162-168). IEEE.
[5] Meng, J., Zhang, J., & Zhao, H. (2012, August). Overview
of the speech recognition technology. In 2012 fourth
international conference on computational and
information sciences (pp. 199-202). IEEE.
[6] Hwang, T. H., Lee, J., Hyun, S. M., & Lee, K. (2020,
October). Implementation of interactive healthcare
advisor model using chatbot and visualization. In 2020
International Conference on Information and
Communication Technology Convergence (ICTC) (pp.
452-455). IEEE.

VOICE RECOGNITION BASED MEDI ASSISTANT

More Related Content

Similar to VOICE RECOGNITION BASED MEDI ASSISTANT (20)

More from IRJET Journal (20)

Recently uploaded (20)

VOICE RECOGNITION BASED MEDI ASSISTANT