Optical Character Recognition
Optical Character Recognition (OCR) aims to automatically convert images of text into machine-readable text, facilitating efficient document processing and information extraction. Current research emphasizes improving OCR accuracy, particularly for challenging scenarios like historical documents, low-resolution images, and complex layouts, often employing transformer-based language models and convolutional neural networks for both character recognition and post-processing error correction. These advancements are crucial for digitizing historical archives, enhancing accessibility to information, and automating various tasks across diverse fields, from document management to scientific literature analysis.
Papers
RoundTripOCR: A Data Generation Technique for Enhancing Post-OCR Error Correction in Low-Resource Devanagari Languages
Harshvivek Kashid, Pushpak Bhattacharyya
Advancing Vehicle Plate Recognition: Multitasking Visual Language Models with VehiclePaliGemma
Nouar AlDahoul, Myles Joshua Toledo Tan, Raghava Reddy Tera, Hezerul Abdul Karim, Chee How Lim, Manish Kumar Mishra, Yasir Zaki
Patchfinder: Leveraging Visual Language Models for Accurate Information Retrieval using Model Uncertainty
Roman Colman, Minh Vu, Manish Bhattarai, Martin Ma, Hari Viswanathan, Daniel O'Malley, Javier E. Santos
OCR Hinders RAG: Evaluating the Cascading Impact of OCR on Retrieval-Augmented Generation
Junyuan Zhang, Qintong Zhang, Bin Wang, Linke Ouyang, Zichen Wen, Ying Li, Ka-Ho Chow, Conghui He, Wentao Zhang