A Review on Building Knowledge Graphs from Scanned Textual Data in Medical Literature for Structured Insights

Abstract

The exponential growth of medical literature presents challenges in extracting, organizing, and utilizing valuable information for healthcare applications. This paper presents a comprehensive review of methodologies for building knowledge graphs (KGs) from scanned medical literature, integrating Optical Character Recognition (OCR) and Natural Language Processing (NLP) techniques. The review addresses the challenges of low-quality scans, domain-specific terminology, data heterogeneity, and entity ambiguity. Furthermore, the paper explores various use cases in healthcare, such as clinical decision support and drug discovery, demonstrating the transformative potential of structured knowledge extraction.