This repository is a curated list of papers, datasets, and resources related to Optical Character Recognition (OCR), particularly focusing on scene text detection and recognition. It serves as a comprehensive reference for researchers and developers working in the field of computer vision and natural language processing who need to understand the state-of-the-art and access relevant materials.

How It Works

The project functions as a meta-resource, aggregating links to academic papers, benchmark datasets, and notable OCR models. It is organized chronologically and by topic, providing a structured overview of the evolution and advancements in OCR technology, from early methods to deep learning approaches. The extensive list of papers and datasets allows users to trace the development of key algorithms and evaluate their performance on standardized benchmarks.

Quick Start & Requirements

This repository is a collection of links and does not have a direct installation or execution command. Users will need to follow the provided links to access papers, datasets, and code repositories.

Highlighted Details

• Extensive collection of research papers spanning from 2009 to 2022, covering a wide range of OCR techniques.
• Comprehensive list of OCR datasets, including ICDAR benchmarks, COCO-Text, SynthText, and others, with descriptions and download links.
• Links to various OCR models and their implementations, such as CRNN, CTPN, EAST, and TextBoxes.

Maintenance & Community

The repository is a static collection of links, with its last update indicated by the latest paper cited. There is no explicit mention of active community engagement or maintenance channels.

Licensing & Compatibility

The repository itself does not host code or datasets, thus it does not have a specific license. Users are advised to check the licenses of the individual linked resources.

Limitations & Caveats

The repository is a curated list and does not provide any executable code or direct access to datasets. Users must navigate external links to obtain the actual resources, and the currency of these links is not guaranteed.