Automated Expiration Date Extraction from Official Documents: A Real-Time OCR-Based Mobile System

Forgetting the expiration dates on official documents can lead to serious problems. These problems can be both legal and financial, and they can also affect daily life. The reminder apps currently in use aim to solve this problem. Margin of error and security vulnerabilities in the management of document validity processes constitute significant risk factors. However, these apps require users to enter information manually. This dependency introduces substantial operational friction, increasing the susceptibility to data entry errors and driving high application abandonment rates. To resolve these issues, the Flutter-based DocGuard mobile application, integrating on-device OCR with a layout-aware heuristic algorithm, was developed to automate and secure data entry, and its performance was evaluated. The developed DocGuard application presents a validated, novel approach that aims to eliminate these risks and prioritizes client-side data privacy. The study’s contribution to the literature consists of three main parts. First, a Layout-Aware Sequence Heuristic (LASH) scoring algorithm capable of reliable date detection across standard structured and semi-structured official documents was designed using the Google ML Kit OCR infrastructure. Second, a privacy architecture prioritizing local processing was established, ensuring closed-loop, strict data confidentiality through standard PBKDF2-HMAC-SHA256 key derivation and AES-256 Client-Side Encryption (CSE) methods optimized for mobile edge devices. The system’s performance was tested using a 5-fold cross-validation method on a balanced dataset consisting of 500 real and open-source documents. The proposed LASH algorithm outperforms standard OCR methods, achieving an exact match accuracy of 94.0% (± 1.2%) for structured documents and 82.0% (± 2.1%) for unstructured documents. Additionally, a statistically significant (p < 0.001) reduction of 70–75% in manual processing time was recorded for users. Although there are some structural limitations in analyzing complex structures and handwriting that require human verification, the findings demonstrate that DocGuard offers a scalable, privacy-preserving digital document lifecycle infrastructure applicable across a wide range of areas, from corporate compliance processes to fintech KYC verifications.

Authors

Institutions

Publication Details

Journal
Gazi University Journal of Science Part A Engineering and Innovation
Published
2026-09-24
DOI
https://doi.org/10.54287/gujsa.1964635
Primary Topic
Handwritten Text Recognition Techniques
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Automated Expiration Date Extraction from Official Documents: A Real-Time OCR-Based Mobile System

Büşra Takgil, Mehmet TAT
Gazi University Journal of Science Part A Engineering and Innovation
Handwritten Text Recognition Techniques
article

Automated Expiration Date Extraction from Official Documents: A Real-Time OCR-Based Mobile System

Büşra Takgil, Mehmet TAT
article en

Abstract

Forgetting the expiration dates on official documents can lead to serious problems. These problems can be both legal and financial, and they can also affect daily life. The reminder apps currently in use aim to solve this problem. Margin of error and security vulnerabilities in the management of document validity processes constitute significant risk factors. However, these apps require users to enter information manually. This dependency introduces substantial operational friction, increasing the susceptibility to data entry errors and driving high application abandonment rates. To resolve these issues, the Flutter-based DocGuard mobile application, integrating on-device OCR with a layout-aware heuristic algorithm, was developed to automate and secure data entry, and its performance was evaluated. The developed DocGuard application presents a validated, novel approach that aims to eliminate these risks and prioritizes client-side data privacy. The study’s contribution to the literature consists of three main parts. First, a Layout-Aware Sequence Heuristic (LASH) scoring algorithm capable of reliable date detection across standard structured and semi-structured official documents was designed using the Google ML Kit OCR infrastructure. Second, a privacy architecture prioritizing local processing was established, ensuring closed-loop, strict data confidentiality through standard PBKDF2-HMAC-SHA256 key derivation and AES-256 Client-Side Encryption (CSE) methods optimized for mobile edge devices. The system’s performance was tested using a 5-fold cross-validation method on a balanced dataset consisting of 500 real and open-source documents. The proposed LASH algorithm outperforms standard OCR methods, achieving an exact match accuracy of 94.0% (± 1.2%) for structured documents and 82.0% (± 2.1%) for unstructured documents. Additionally, a statistically significant (p < 0.001) reduction of 70–75% in manual processing time was recorded for users. Although there are some structural limitations in analyzing complex structures and handwriting that require human verification, the findings demonstrate that DocGuard offers a scalable, privacy-preserving digital document lifecycle infrastructure applicable across a wide range of areas, from corporate compliance processes to fintech KYC verifications.

Gazi University Journal of Science Part A Engineering and Innovation(Advanced Online Publication)
Düzce Üniversitesi (TR)
Industry, innovation and infrastructure
Openalex Percentile: Top 14%
Handwritten Text Recognition Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.