Valid: Fraud detection mobile application for government identity cards using optical character recognition and SHA-256 algorithm for data application

ABSTRACT: Government-issued identification (ID) cards are essential for verifying individuals identities and granting access to various services. However, the increasing prevalence of ID-related fraud poses serious security challenges. This study presents ValID: A Fraud Detection Mobile Application for Government Identity Cards, which integrates Optical Character Recognition (OCR) for accurate text extraction and the Secure Hash Algorithm-256 (SHA-256) for preserving data integrity. Developed using the Agile Software Development Life Cycle (SDLC), the system showed 100% OCR accuracy in normal scanning conditions, ensuring reliable text recognition. Yet, under difficult scenarios such as handwritten, smudge, or blurry ID, the system’s OCR failed to extract data effectively, resulting in 0% accuracy. It also struggled with varying font styles and sizes, producing an average Character Error Rate (CER) of 64.25%, indicating a need for further enhancement in processing diverse formats. To support transparency and accountability, ValID includes an audit log feature that records all transactions. Meanwhile, SHA-256 ensures data protection by converting user inputs, like passwords, into complex hash values. For instance, a basic password such as “1234” is transformed into an irreversible alphanumeric string, safeguarding it from unauthorized access and tampering. Despite its promising results, ValID still has room for improvement. Future enhancements may include incorporating convolutional neural networks (CNNs) for more effective fraud detection, enabling real-time alerts for suspicious activity, and adding features like multi-factor authentication (MFA). Overall, ValID delivers a secure and efficient tool for combating government ID fraud, reinforcing trust in identity verification systems.

Filipiniana