Methodology for the Semantic Retrieval of Documents using Workflows based on RAG and OCR

Torres-Cabrera J.F.; Auccahuasi W. · International Conference on Intelligent and Sustainable AI Systems, ICOSAAS 2026 - Conference Proceedings (2026) · DOI: 10.1109/icosaas68663.2026.11648790

Resumen

Currently, many physical documents contain important information that cannot be incinerated, hence the need to continue accessing this valuable information. This article presents a methodology for exploiting the information contained in physical documents through digitization and image processing. This is achieved by combining different technologies and libraries, such as OCR and RAG techniques, to query the information contained in the document. The methodology's application was tested by analyzing information contained in academic documents, which allows the methodology to be scaled and replicated, thanks to the detailed explanation of the solution's implementation.

Instituciones: Universidad Privada del Norte