Optical Character Recognition (OCR) plays a pivotal role in digitising historical and regional documents, enabling preservation and accessibility of cultural heritage. This paper presents an Enhanced OCR Pipeline that addresses the challenges of processing degraded, handwritten, and multi-lingual regional records. The proposed system integrates advanced image preprocessing techniques including CLAHE contrast enhancement, bilateral filtering, and Sauvola binarization with intelligent layout analysis, table detection, and multi-format export capabilities. Built on the Tesseract OCR engine, the pipeline supports over 100 languages including RTL scripts and provides per-word confidence scoring with comprehensive quality metrics. Experimental results demonstrate an accuracy of 95%+ on clean modern documents and 85%+ on historical records.
Paper
The full text of this publication is not hosted on 44B due to licensing.
Read it at OpenAlex