OTOBOOK: Automated Library Cataloging System Based on OCR, AI, and RPA
Background and Urgency
In the rapidly evolving digital era, libraries as information-providing institutions face significant challenges in managing increasingly growing collections. The East Java Provincial Library, as one of the leading educational and research institutions, holds crucial responsibility in providing relevant, accurate, and easily accessible information to the community.
However, reality shows that traditional cataloging processes still conducted manually have become the main bottleneck in modern library operations. Empirical data reveals cataloging error rates reaching 1.42%, with a more concerning figure that 64.2% of these errors went undetected or unresolved. This situation not only impacts library service quality but also creates disproportionate workloads for librarians who must handle continuously increasing work volumes.

The annual increase in published books, both in print and digital formats, further exacerbates this situation. Librarians face time pressure to process new collections while maintaining high cataloging quality standards. This condition clearly demonstrates the urgent need to adopt technological solutions that can automate cataloging processes without sacrificing metadata accuracy and quality.
mats and image qualities but have also been optimized to handle specific characteristics of Indonesian publications, including mixed-language usage (Indonesian-English), diverse writing formats, and varying print qualities.

OCR advantages in OTOBOOK include:

Adaptive preprocessing capabilities handling various lighting conditions and scan qualities
Automatic correction algorithms to overcome geometric distortions
Multi-language processing support for bilingual or multilingual publications
Batch processing capability for simultaneous multiple document processing
Artificial Intelligence (AI) Component
The AI ​​module in OTOBOOK employs machine learning and natural language processing approaches to analyze and classify metadata with high accuracy. The AI ​​system has been trained using comprehensive datasets covering various genres, subjects, and publication formats commonly found in Indonesian libraries.

Integrated AI features include:

Automatic subject classification based on Dewey Decimal Classification (DDC)
Genre detection and content categorization
Author name disambiguation and standardization
Publication information extraction and verification
Quality assurance automated checking
Robotic Process Automation (RPA)
The RPA component functions as an orchestrator that automates the entire cataloging workflow, from data input to integration with national catalog systems. RPA ensures process consistency and reduces human errors that frequently occur in manual input.

RPA capabilities in OTOBOOK:

Automated data entry to various catalog systems
Cross-platform integration with SLiMS, OPAC, and national catalog systems
Automated backup and version control
Exception handling and error recovery mechanisms
Comprehensive audit trail for tracking and compliance