Search Results for "structure"

DeepSeek-OCR-2 is the second-generation optical character recognition system developed to improve document understanding by introducing a “visual causal flow” mechanism, enabling the encoder to reorder visual tokens in a way that better reflects semantic structure rather than strict raster scan order. It is designed to handle complex layouts and noisy documents by giving the model causal reasoning capabilities that mimic human visual scanning behavior, enhancing OCR performance on documents with rich spatial structure. The repository provides model code and inference scripts that let researchers and developers run and benchmark the system on both images and PDFs, with support for batch evaluation and optimized pipelines leveraging vLLM and transformers.

dots.ocr is a cutting-edge multilingual document parsing system built on a unified vision-language model that combines layout detection, text recognition, and structural understanding into a single architecture. Unlike traditional OCR pipelines that rely on multiple specialized components, dots.ocr integrates these processes end-to-end, reducing error propagation and improving consistency across tasks. The model is designed to recognize virtually any human script, making it highly effective...

WebDjVuTextEd allows to edit the text layer of OCR'ed DjVu documents in a web browser. You can modify the structure (paragraphs, lines, words...) create, delete, edit text nodes, modify their container box by mouse, and run a spellchecker. The program does not directly read the DjVu files, it requires exported XML text data and images. When using without a webserver, you can open and save local files, but cannot take advantages of auto-save and spell checking.

Note as of 2013-09-13: I'm moving this project over to github due to this: http://www.gluster.org/2013/08/how-far-the-once-mighty-sourceforge-has-fallen/ Feel free to rejoin the more updated versions on https://github.com/mnott/PDFOCRWrapper Thanks. Matthias -- This is a wrapper written in Java that allows to recursively iterate a directory structure and call an OCR engine on each found PDF on the condition that it hat not yet been called for that PDF. It works well with the ABBYY OCR Engine for Linux.

Unlimited-OCR is Baidu’s open-source optical character recognition (OCR) model designed to accurately extract and understand text from complex documents, images, and multilingual content. Unlike traditional OCR systems that focus only on text detection and transcription, Unlimited-OCR combines advanced document parsing with language understanding, enabling it to recognize structured elements such as tables, formulas, charts, and mixed-layout documents while preserving their logical...