KhwajaFareed University of Engineering & Information

Technology, Rahim Yar Khan

Department of Information Technology

Bachelor of Science in Information Technology

COSC-4301
Final Year Project Proposal
AI Health Spectrum

Academic Year 2021-2025

Sr # Student Name Reg. No.

01 Mubashra Bibi INFT212101020

Page | 1
Contents
1. Abstract...................................................................................................................................3
2. Introduction and Background..............................................................................................3
3. Problem Statement.................................................................................................................4
1. Literature Review..................................................................................................................4
2. Project Rationale....................................................................................................................5
3. Aims and Objectives..............................................................................................................5
4. Scope of the Project................................................................................................................5
5. Proposed Methodology..........................................................................................................6
6. Proposed solution and anticipated results...........................................................................7
Anticipated Results...................................................................................................................7
7. Gantt chart.............................................................................................................................8
11. References............................................................................................................................9

Page | 2
1. Abstract
In the era of digital healthcare, the demand for accessible and accurate medical assistance is paramount.
This project presents a comprehensive web application designed to fulfill multiple healthcare needs
seamlessly. The platform incorporates advanced artificial intelligence (AI) models for disease prediction,
deep learning modules for pneumonia detection through X-ray images, and a user-friendly interface for
locating doctors based on various parameters.
The core feature of the platform is its AI model, capable of predicting 42 different diseases based on 132
symptoms input by the user. Upon diagnosis, the system provides tailored recommendations including
disease information, exercises, home remedies, and precautions. Additionally, users have the option to
download detailed reports summarizing their diagnosis and recommendations, enabling them to share vital
information with their healthcare providers for further evaluation and guidance. This integration of
technology and personalized care ensures that users receive comprehensive support throughout their
healthcare journey, empowering them to make informed decisions and take proactive steps toward better
health outcomes.
The platform offers a specialized module for pneumonia detection using deep learning algorithms trained on
X-ray images. This module aids in rapid and accurate diagnosis of pneumonia, facilitating timely
intervention and treatment. The integration of this module not only streamlines the diagnostic process but
also enhances patient outcomes by expediting the delivery of appropriate care. Rapid detection of
pneumonia allows for prompt initiation of treatment, reducing the risk of complications and improving
overall prognosis. Moreover, by facilitating early identification of the condition, the platform empowers
healthcare professionals to implement targeted interventions tailored to the individual needs of each patient,
thus optimizing treatment efficacy.
Furthermore, the platform includes a user-friendly doctor search functionality. Users can easily locate
doctors based on their location, specialization, or name, providing convenience in accessing healthcare
services. Doctors' details, including their qualifications, experience, and contact information, are readily
available to aid in informed decision-making.

2. Introduction and Background

The proposed project aims to develop an integrated AI Health Spectrum web application that leverages
advanced technologies such as artificial intelligence (AI) and deep learning to provide accessible and
accurate medical assistance. In the digital era, the demand for such solutions is ever-growing, and this
project seeks to address this need by offering a comprehensive platform that caters to various healthcare
requirements seamlessly. The core feature of the platform is its AI model, capable of predicting 42 different
diseases based on 132 symptoms input by the user. Upon diagnosis, the system provides tailored
recommendations including disease information, exercises, home remedies, and precautions. Additionally,
users have the option to download detailed reports summarizing their diagnosis and recommendations,
enabling them to share vital information with their healthcare providers for further evaluation and guidance.
This integration of technology and personalized care ensures that users receive comprehensive support
throughout their healthcare journey, empowering them to make informed decisions and take proactive steps
toward better health outcomes.

Noticing a growing demand for easily accessible, precise, and prompt medical assistance, there is a
challenge in obtaining reliable health details, understanding symptoms, and receiving appropriate care
without delays. Additionally, the conventional healthcare system often lacks the flexibility and personalized
touch necessary to efficiently address diverse patient needs. To tackle these challenges, the vision involves
developing a groundbreaking web application that harnesses cutting-edge technologies to provide
Page | 3
comprehensive healthcare solutions. The project's objective is to empower individuals through a user-
friendly platform, facilitating personalized medical assistance, access to reliable information, and seamless
connections with healthcare professionals.

3. Problem Statement
The user has identified a significant challenge in accessing reliable and timely medical assistance. Many
people struggle to find accurate health information and connect with healthcare providers efficiently. To
address this, a pioneering web application is being developed. This project aims to provide a simple and
accessible platform for users to receive personalized medical guidance, access trustworthy information, and
easily connect with healthcare professionals. By leveraging advanced technologies, such as AI and deep
learning, the platform will offer tailored solutions to individual health concerns. Through this initiative,
users will be empowered to take control of their health journey, leading to better outcomes and improved
well-being.

4. Literature Review
Lack of diagnosis and treatment kills most people today. These patients have problems using or accessing
medical care rapidly. This may be useful in remote places with minimal medical resources or urgent
patients. Thus, we created a self-assessment software that predicts 40 ailments using machine learning and
symptoms. Correct model predictions are needed for self-assessment. Machine learning approaches have
accuracy issues, which is critical in such situations. We recommend a technique to diagnose diseases from
symptoms. Data was categorized using machine learning models such K Nearest Neighbor, Support Vector
Machine, Decision Tree, Random Forest Classifier, Naïve Bayes, and Artificial Neural Network after
preprocessing. All model probability vectors are totaled and separated by model number after hyper
parameter optimization. The threshold parameters define the application screen's ailment count. The
programme displays illness predictions from the model. Our model outperforms others [1].

Annually, pneumonia kills many youngsters. 880,000 children under 5 died from pneumonia in 2016, with
1.2 million cases recorded. Thus, pneumonia kills many youngsters in South Asia and Sub-Saharan Africa.
Even in affluent nations like the US, pneumonia remains a top 10 killer. Early identification and treatment
of pneumonia may considerably decrease child mortality in high-prevalence nations. Using x-ray pictures,
this article provides Convolutional Neural Network models to identify pneumonia. Multiple Convolutional
Neural Networks were trained to categories x-ray pictures into pneumonia and non-pneumonia by
modifying parameters, hyperparameters, and convolutional layers. 6 models are discussed in the study. First
and second models have two and three convolutional layers. Another four pre-trained models are VGG16,
VGG19, ResNet50, and Inception-v3. First and second models demonstrate 85.26% and 92.31% validation
accuracy. VGG16, VGG19, ResNet50, and Inception-v3 had 87.28%, 88.46%, 77.56%, and 70.99%
accuracy [2].

A rapidly developing, diversified medical analytic discipline is Computer Aided Diagnosis (CAD). Recent
attempts have been made to build computer-aided diagnostic applications since medical diagnosing errors
might lead to false treatments. ML is crucial to computer-assisted diagnosis. Body organs cannot be
identified using a simple equation. Thus, pattern recognition needs instance-based training. In biomedicine,
pattern recognition and ML promise to enhance illness identification and approach. Dispassionate decision-
Page | 4
making is also respected. ML is a good way to create automated algorithms for high-dimensional and multi-
modal biomedical data. This survey report compares ML algorithms for heart and diabetes illness diagnosis.
It emphasizes ML algorithms and methods for illness diagnosis and decision-making [3].

5. Project Rationale
The project is driven by the urgent need to provide individuals with accessible, accurate, and personalized
healthcare solutions. Leveraging advanced technologies such as AI and deep learning, the web application
aims to enhance diagnostic accuracy, particularly in conditions like pneumonia, while also empowering
users to actively manage their health. By offering personalized recommendations based on user-input
symptoms and facilitating seamless communication with healthcare providers, the project seeks to bridge
gaps in healthcare access and delivery. Ultimately, the goal is to transform the healthcare experience,
making it more convenient, efficient, and tailored to individual needs.

6. Aims and Objectives

The overarching goal of the project is to revolutionize the AI Health Spectrum experience by developing a
comprehensive web application that seamlessly integrates advanced technologies and user-centric design
principles. The main focused objectives of the project include:

1) Improving Diagnostic Accuracy: To develop AI-driven models capable of accurately predicting a wide
range of diseases based on user-input symptoms, thereby facilitating early detection and timely
intervention.
2) Streamlining Pneumonia Detection: To implement deep learning algorithms for the rapid and accurate
detection of pneumonia from X-ray images, enabling prompt diagnosis and treatment initiation.
3) Empowering Users with Personalized Care: To offer personalized recommendations tailored to
individual health needs, including disease information, exercises, home remedies, and precautions.
4) Facilitating Collaboration with Healthcare Providers: To enable search fine there doctor to
communication and users, ensuring continuity of care and optimized treatment outcomes.
5) Enhancing Healthcare Accessibility: To provide individuals with easy access to reliable healthcare
information and services, irrespective of their geographic location or socioeconomic status.
6) Ensuring User-Friendly Experience: To design an intuitive and user-friendly interface that prioritizes
simplicity, accessibility, and ease of navigation, catering to the diverse needs of users.

7. Scope of the Project

The scope of the project encompasses the development of a comprehensive web application focused on
providing accessible and accurate AI Health Spectrum solutions. Key deliverables include the
implementation of AI-driven disease prediction models, deep learning algorithms for pneumonia detection
from X-ray images, and a user-friendly interface for symptom input and accessing AI Health Spectrum
resources. Deadlines will be set for each phase of development, ensuring timely delivery of features and
functionalities. Costs will be allocated for software development, data acquisition, and infrastructure setup
to meet project objectives within budget constraints.

Page | 5
8. Proposed Methodology
The development of the envisioned AI Health Spectrum web application will follow a systematic and
iterative approach, encompassing the following steps:

1) Requirement Analysis: Conduct a thorough analysis of project requirements, dataset for detecting
disease base on symptoms and chest x-ray dataset for pneumonia detecting, also including the need
for searching doctors by name, location, education, disease, and specialization, as well as
incorporating user feedback mechanisms.
2) Design Phase: Create detailed specifications for the web application's architecture, database schema,
user interface, and the addition of search functionalities for doctors.
3) Data Acquisition and Preprocessing: Gather and preprocess data related to doctors, including their
names, locations, education, specialties, and diseases they treat.
4) Analyze User Input: Develop algorithms for analyzing user input and symptoms, processing them,
and predicting the disease based on symptoms.
5) Pneumonia Detection: Develop AI models to analyze chest X-ray images uploaded by users,
detecting pneumonia, and providing relevant output.
6) Mechanisms for searching: Implement mechanisms for searching doctors based on specified
criteria, such as name, location, education, disease, and specialization.
7) User Interface Implementation: Design and develop user-friendly interfaces for searching doctors
based on various criteria, providing users with easy navigation and access to relevant AI Health
Spectrum professionals.
8) Details and Report: Provide users with access to detailed information about diseases and their
predictions. Users should also be able to download reports summarizing their diagnosis and
recommendations for further evaluation and guidance.
9) Admin Control: Implement functionality for admin management, allowing the admin to manage the
entire website, including handling user feedback. Additionally, the admin can add more doctors and
their details to the system.
10) Integration and Testing: Integrate the doctor search functionalities into the web application,
ensuring seamless interaction with other components. Conduct rigorous testing to verify the
accuracy and usability of the search feature.
11) Deployment and Evaluation: Deploy the updated web application with doctor search capabilities,
monitor user interactions, and gather feedback on the usability and effectiveness of the search
functionalities.

Page | 6
 12) User Feedback Mechanisms: Implement mechanisms for collecting user feedback, such as ratings,
and comments, to continuously improve the web application based on user input.

Figure 1: Process Model Diagram

9. Proposed solution and anticipated results

The projected system gets to modernize AI Health Spectrum delivery by integrating advanced technologies
such as AI-driven disease prediction and deep learning algorithms for pneumonia detection. Through
intuitive user interfaces, users can input symptoms and receive accurate disease predictions, along with
tailored recommendations for further action. Additionally, the system offers doctor search functionality,
enabling users to find healthcare professionals based on various criteria, streamlining access to relevant
medical expertise. By incorporating user feedback mechanisms, continuous improvements ensure a user-
centric approach to development, enhancing the overall effectiveness and usability of the AI Health
Spectrum solution.

Anticipated Results:
The anticipated results of the proposed solution are multifaceted, aiming to significantly improve AI Health
Spectrum accessibility, diagnostic accuracy, and overall user experience. With accurate disease predictions
and pneumonia detection capabilities, individuals can receive timely interventions and personalized care,
leading to better health outcomes. Moreover, the integration of doctor search functionalities ensures easier
access to AI Health Spectrum professionals, particularly beneficial for underserved communities. User
feedback mechanisms facilitate ongoing improvements, ensuring the system remains responsive to user
needs and preferences, ultimately revolutionizing the AI Health Spectrum landscape.

Page | 7
10.Gantt chart

Page | 8
 References
[1] .A. Sridhar, A. Mawia, and A. L. Amutha, “Mobile Application Development for Disease Diagnosis based
on Symptoms using Machine Learning Techniques,” Procedia Comput. Sci., vol. 218, pp. 2594–2603, Jan.
2023, doi: 10.1016/j.procs.2023.01.233.
[2] R. Jain, P. Nagrath, G. Kataria, V. Sirish Kaushik, and D. Jude Hemanth, “Pneumonia detection in chest X-
ray images using convolutional neural networks and transfer learning,” Measurement, vol. 165, p. 108046,
Dec. 2020, doi: 10.1016/j.measurement.2020.108046.
[3] P. Hamsagayathri and S. Vigneshwaran, “Symptoms Based Disease Prediction Using Machine Learning
Techniques,” in 2021 Third International Conference on Intelligent Communication Technologies and
Virtual Mobile Networks (ICICV), Feb. 2021, pp. 747–752. doi: 10.1109/ICICV50876.2021.9388603.
`

Page | 9