This project is mca students final year project report

1. AI-Driven Medical Fundraising Verification
System to Detect and Prevent Fraudulent
Treatment Requests

2. OUTLINE
1. Introduction
2. Objective
3. Literature Review
4. Proposed Model
5. Application of Proposed System
6. Sample Code
7. Screen Layout
8. Result and Conclusion
9. References

3. Introduction
Scammers create fake treatment bills and manipulate social media or crowdfunding platforms to
deceive donors, reducing trust in genuine medical fundraising.
Traditional fraud detection methods rely on human review, which is slow, inefficient, and often
fails to detect sophisticated fraudulent attempts.
Implementing an automated system enhances security by using advanced AI techniques to verify
medical fund requests efficiently and accurately.
The system utilizes YOLOv8 for detecting text in medical bills and PaddleOCR for extracting and
recognizing patient and hospital details.
Extracted text is compared with trusted hospital records using the Fuzzy Matching Algorithm to
detect inconsistencies and block fraudulent requests.

4. Objective
The aim of this project is to develop an AI-powered system for detecting and preventing
fraudulent medical fund requests by verifying treatment documents using advanced text
detection, recognition, and pattern-matching techniques.
Objective
To automate the verification of medical fund requests using AI-based fraud detection.
To integrate YOLOv8 for detecting text in treatment bills and PaddleOCR for text
recognition.
To apply Fuzzy Matching Algorithm to compare extracted text with authentic hospital
records.
To enhance donor trust by ensuring that only genuine medical fund requests are approved.
To minimize fraudulent activities in medical crowdfunding by providing a reliable and
efficient verification system.

5. 1.IHSAN: A Secure and Transparent Crowdfunding Platform Leveraging
Comprehensive Decentralized Technologies
(Omar Khalid Alia; Duaa Mohammad Suleiman, 2024)
Methodology:
IHSAN utilizes blockchain technology and BigchainDB to build a decentralized
crowdfunding platform that ensures trust, transparency, security, and scalability.
Algorithms:
Blockchain enables decentralized control and immutable transactions, ensuring
transparent and accountable fund allocation.
Findings:
IHSAN reduces platform fees to just 1% (compared to 8% and 2.9% in traditional
platforms), cuts transaction times to about 2 minutes, and allows unrestricted global
access—enhancing donor trust and increasing project success rates.
LITERATURE SURVEY

6. 2.How Do Project Updates Influence Fundraising on Online Medical
Crowdfunding Platforms? Examining the Dynamics of Content Updates
(Yi Wu; Miao Zhang; Yi Shen, 2024)
Methodology:
Analyzed 2,334 medical crowdfunding projects (Jan 2020–July 2021) using
Aristotle’s persuasion theory to study the impact of update types on
fundraising.
Algorithms:
Updates were classified as credibility, rational, or emotional appeals and
evaluated through statistical analysis.
Findings:
All appeal types boost fundraising. Credibility appeals lose effectiveness over
time, while rational and emotional appeals grow more impactful, showing the
dynamic role of updates in fundraising success.
LITERATURE SURVEY

7. 3.A Blockchain-Based Crowdsourcing Loan Platform for Funding
Higher Education in Developing Countries
(Kwame Omono Asamoah; Adjei Peter Darko, 2023)
Methodology:
Proposes a blockchain-based decentralized platform connecting students
with investors through financial institutions to fund higher education.
Algorithms:
Uses blockchain for security, smart contracts for loan agreements, and a
decentralized ledger for transparent transactions.
Findings:
Reduces government financial strain, offers investors interest returns,
ensures secure transactions, and improves students' access to education
funding.
LITERATURE SURVEY

8. Proposed Model
The proposed system aims to enhance the detection and prevention of fraudulent medical
fund requests by integrating AI-driven technologies. It automates the verification process,
ensuring accuracy and efficiency while minimizing human intervention.
AI-Based Fraud Detection
The system employs YOLOv8 for detecting text regions in medical bills and PaddleOCR
for extracting textual information such as hospital names, patient details, and treatment
costs. These extracted details are then analyzed to identify potential discrepancies.
Pattern Matching for Verification
To ensure authenticity, the system utilizes the Fuzzy Matching Algorithm, which compares
extracted text with a trusted hospital dataset. This method effectively measures similarity
and detects inconsistencies in treatment details, preventing fraudulent fund requests.

9. Proposed Model
Automated Document Processing
Unlike traditional manual verification methods, the proposed system automates
document processing, significantly reducing the time required for fraud detection. It
eliminates human errors and ensures consistency in identifying fake medical fund
requests.
Secure and Transparent Donation Process
The system enhances donor confidence by providing a transparent verification process.
Only verified medical fund requests are displayed to potential donors, ensuring that
contributions reach genuine beneficiaries.

10. Proposed Model
Detects and blocks fraudulent medical fund requests using AI.
Automates verification, reducing manual effort and errors.
Builds donor trust through transparent validation.
Enables real-time processing for quick fraud detection.
Secures transactions by verifying fund requests before approval.
Supports scalability for hospitals and crowdfunding platforms.
Lowers operational costs by minimizing manual fraud detection.
Provides an easy-to-use interface for patients and donors.

11. System Architecture
AI Driven Fraud Detector
Preprocessing
Register
View Flagged Request
User management
Login
Text Recognition
Text Region Detection
Extract Upload Files Documents
Donor
Admin
Fund Requester
Medical Fund Fraud Detector Web
App
View Fund Request
View Donor Responses
Login
Post Fund Request
Upload Documents
Approve/Block Req
Receive Donation
Notification
Register
Login
View Genuine Req.
Donate Payment
Notification
Hospital Database
Pattern Matching
Fund Request Verification

12. Applications of Proposed Model
1. Medical Fund Fraud Detector Web App
2. End User
2.1. Admin
2.2. Patient/Fund Requester
2.3. Donors
3. Hospital Database Integrator
4. Medical Fund Request
5. Fraud Detection
5.1. Preprocessing
5.2. Text Region Detection
5.3. Text Recognition
5.4. Pattern Matching
6. Fund Request Verification
7. Donor Payment Processing
8. Notification

13. 1. Medical Fund Fraud Detector Web App
This web application is designed to detect fraudulent medical fund requests by
leveraging Machine Learning and Computer Vision.
It integrates Python, Flask, MySQL, Wampserver, TensorFlow, Pandas, Scikit-Learn,
Matplotlib, NumPy, Seaborn, Pillow, OpenCV, and Bootstrap to provide a robust and
efficient system.
It aims to prevent fraudulent medical fund requests by utilizing YOLOv8-based object
detection and pattern-matching techniques.
The platform connects patients, donors, and administrators to ensure funds are distributed
only to genuine applicants.

14. 2. End User
This module defines the user roles and functionalities.
2.1. Admin
Login & Authentication – Secure login access to the system.
User Management – Manage users (Fund Requesters and Donors).
View Fund Requests & Responses – Monitor all fund requests and their verification
status.
View Flagged Requests – Identify suspicious fund requests flagged as fraudulent.

15. 2.2 Patient / Fund Requester
Register & Login – Create an account and log in securely.
Post Fund Request – Upload medical documents and request financial assistance.
View Verified Result – Check the approval or rejection status of the request.
Receive Payment – If the request is genuine, receive donations.
2.3 Fund Donor
Register & Login – Sign up and log in securely.
View Genuine Requests – Browse and verify fund requests marked as genuine.
Donate Payment – Make donations securely to verified patients.

16. 3. Hospital Database Integrator
Integrates with hospital records and medical billing databases to cross-verify patient
details.
Fetches medical records, prescriptions, and hospital authentication stamps to validate
requests.
Helps in identifying discrepancies between submitted documents and hospital records.

17. 4. Medical Fund Request
Collects fund request details, including patient information, treatment details, hospital
bills, and medical reports.
Stores requests in a centralized database for validation.
Provides document upload functionality for medical certificates and treatment invoices.

18. 5. Fraud Detection
This module applies Machine Learning and Computer Vision techniques to detect
fraud in medical fund requests.
5.1 Preprocessing
Converts images to grayscale and resizes them for uniformity.
Applies Noise Filters (Gabor Filter / Median Filter) to remove distortions.
Binarizes text to enhance clarity for further processing.
5.2 Text Region Detection
Uses OpenCV and YOLOv8 to detect text areas in scanned documents.
Identifies stamps, signatures, and printed text regions to verify authenticity.

19. 5.3 Text Recognition
Extracts text from detected regions using OCR (Tesseract / EasyOCR).
Recognizes hospital names, patient details, and billing amounts.
5.4 Pattern Matching
Compares extracted text with hospital database records to verify legitimacy.
Detects forged hospital stamps, altered signatures, and duplicated requests.
Flags requests with mismatched or manipulated information.

20. 6. Fund Request Verification
The system classifies requests into three categories:
oApproved – Verified and ready for donor support.
oFlagged – Suspicious and requires manual admin review.
oRejected – Identified as fraudulent and removed from the system.
Genuine requests are made visible to donors, while flagged requests require further
verification.

21. 7. Donor Payment Processing
Provides a secure payment gateway for donors to contribute.
Ensures funds are transferred only to verified patients.
Tracks donation transactions and receipts for transparency.

22. 8. Notification
Sends real-time alerts and updates to all users.
Admin Alerts – Notifies about flagged fund requests.
Patient Alerts – Updates on fund request approval, verification, and payments received.
Donor Alerts – Confirmation of successful donations and updates on fund utilization.

23. Source Code
Login
def login():
msg=""
if request.method=='POST':
uname=request.form['uname']
pwd=request.form['pass']
cursor = mydb.cursor()
cursor.execute('SELECT * FROM mf_user WHERE uname = %s AND pass = %s', (uname, pwd))
account = cursor.fetchone()
if account:
session['username'] = uname
return redirect(url_for('userhome'))
else:
msg = 'Incorrect username/password!'

24. Source Code
def reg_user():
msg=""
if request.method=='POST':
name=request.form['name']
mobile=request.form['mobile']
email=request.form['email']
address=request.form['address']
city=request.form['city']
acc_name=request.form['acc_name']
bank_name=request.form['bank_name']
'uname=request.form['uname']
pass1=request.form['pass']
now = datetime.datetime.now()
rdate=now.strftime("%d-%m-%Y")
mycursor = mydb.cursor()
bank_name=request.form['bank_name']
account=request.form['account']
branch=request.form['branch']
gpay_number=request.form['gpay_number']
uname=request.form['uname']
pass1=request.form['pass']
now = datetime.datetime.now()
rdate=now.strftime("%d-%m-%Y")

25. SCREEN LAYOUTS

26. SCREEN SHOTS

27. SCREEN SHOTS

28. SCREEN SHOTS

29. SCREEN SHOTS

30. SCREEN SHOTS

31. SCREEN SHOTS

32. SCREEN SHOTS

33. Result and Conclusion
In conclusion, the project enhances transparency and security in medical fund requests by
leveraging AI and machine learning to detect fraudulent claims. Using YOLOv8, OCR,
and pattern matching, it identifies forged bills, manipulated documents, and fake hospital
stamps with high accuracy.Key functionalities include fraud detection, automated
verification, secure donor transactions, and real-time notifications. By integrating
hospital databases, the system streamlines the donation process and prevents financial
exploitation.This project fosters trust in medical crowdfunding and paves the way for
future enhancements such as blockchain integration and advanced AI-based fraud
detection.

34. References
1. A. Fong, M. Jain, W. Sacks, A. Ho and Y. Chen, "Crowdfunding campaigns and thyroid surgery: Who what
where and how much?", J. Surg. Res., vol. 253, pp. 63-68, Sep. 2020.
2. Z. Ba, Y. Zhao, S. Song and Q. Zhu, "Understanding the determinants of online medical crowdfunding
project success in China", Inf. Process. Manage., vol. 58, Mar. 2021.
3. J. Mejia, G. Urrea and A. J. Pedraza-Martinez, "Operational transparency on crowdfunding platforms: Effect
on donations for emergency response", Prod. Oper. Manage., vol. 28, pp. 1773-1791, Jul. 2019.
4. K. Zhao, L. Zhou and X. Zhao, "Multi-modal emotion expression and online charity crowdfunding success",
Decis. Support Syst., vol. 163, 2022.
5. J.-R. Hou, J. Zhang and K. Zhang, "Pictures that are worth a thousand donations: How emotions in project
images drive the success of online charity fundraising campaigns? An image design perspective", MIS
Quart., vol. 47, pp. 535-584, 2023.