Welcome to the newly upgraded BRAC University Institutional Repository! Following our recent system upgrade, we are actively organizing our collections. While the category counters on the homepage are currently syncing and may temporarily display low numbers, rest assured that our full repository of over 27,000 items remains safely intact. Please use the search bar above to easily access all scholarly outputs, theses, and institutional documents while we complete this categorization process.

Implementation of Bengali voice signature authentication techniques in financial systems using deep learning

dc.contributor.advisorSadeque, Farig Yousuf
dc.contributor.authorRashid, Mahinur
dc.contributor.authorAhmad, Safwan
dc.contributor.authorFaiza Binte, Arif
dc.contributor.authorAnjum, Ramisa
dc.contributor.authorRimu, Sanjida Arefin
dc.contributor.departmentDepartment of Computer Science and Engineering
dc.date.accessioned2026-02-17T04:24:00Z
dc.date.available2026-02-17T04:24:00Z
dc.date.copyright2025
dc.date.issued2025-10
dc.descriptionCataloged from PDF version of thesis.
dc.descriptionIncludes bibliographical references (pages 128-131).
dc.descriptionThis thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2025.en_US
dc.description.abstractIn developing-nations like Bangladesh, the need for more secure user-friendly authentication methods to protect financial transactions are a substantial need nowadays due to the growing digitalization in finances. Nowadays the need for more secure, fast and user-friendly authentication methods in the financial sectors is increasing rapidly and here the voice-based authentication provides an aspiring substitution to the regular conventional methods. The aim of this paper is to discuss the area of coverage for the potential application of deep learning models in reference to implementing voice signature authentication procedures for cash-out processes in physical banking and digital mobile payment systems. The methodology includes the development and testing of a voice verification system and its integration with the already existing banking and mobile payment infrastructures. Initially, the Bengali voice data samples are collected from available sources, and then analyzed using statistical methods based on differentiable voice/tone metrics and thematic analysis for purposes of user identification and detection of fraud and AI generated voices. It greatly relies on the use of deep learning techniques in voice signature analysis and authentication along with ensuring the authentication procedures are safe and smooth. Key findings support this study in regarding voice signature authentication process, as it can be a promising and cheap addition to the security protocols by increasing accuracy, reliability and also reduce fraud along with improving smooth user experience in cash-out processes. Additionally, this study offers real-world proof of the successful use of deep learning in financial authentication systems, this study adds up the study to the vast domain of deep learning. This study intends to contribute to the improvement of financial security methods through preserving the sensitive financial data and by tackling the growing need for more secure and user-friendly intuitive authentication techniques. Subsequent future research could dive deeper into voice verification technology and its application in other sensitive sectors for further developments; opening a new door of authentication protocol in a country like Bangladesh.en_US
dc.description.degreeBachelor of Science in Computer Science and Engineering
dc.description.statementofresponsibilityMahinur Rashid
dc.description.statementofresponsibilitySafwan Ahmad
dc.description.statementofresponsibilityFaiza Binte Arif
dc.description.statementofresponsibilityRamisa Anjum
dc.description.statementofresponsibilitySanjida Arefin Rimu
dc.format.extent131 pages
dc.identifier.otherID 22301714
dc.identifier.otherID 21101311
dc.identifier.otherID 21201498
dc.identifier.otherID 21301399
dc.identifier.otherID 21201655
dc.identifier.urihttp://hdl.handle.net/10361/27531
dc.language.isoenen_US
dc.publisherBRAC Universityen_US
dc.rightsBRAC University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.
dc.subjectBengali voice authenticationen_US
dc.subjectDeep learningen_US
dc.subjectFinancial securityen_US
dc.subjectMobile payment systemsen_US
dc.subjectFraud detectionen_US
dc.subjectUser-friendly authenticationen_US
dc.subject.lcshSpeaker recognition.
dc.subject.lcshVoiceprints.
dc.subject.lcshBiometric identification.
dc.subject.lcshAuthentication (Computer security).
dc.subject.lcshDeep learning (Machine learning).
dc.subject.lcshElectronic funds transfers.
dc.subject.lcshBengali language--Speech recognition.
dc.subject.lcshAutomatic speech recognition.
dc.titleImplementation of Bengali voice signature authentication techniques in financial systems using deep learningen_US
dc.typeThesisen_US

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
21101311, 22301714, 21201498, 21301399, 21201655_CSE.pdf
Size:
4.89 MB
Format:
Adobe Portable Document Format
Description:

License bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description: