awesome-deep-text-detection-recognition

Why we included this project

Anyone building or researching OCR systems will find this list a fast way into the deep-learning literature on scene text. Papers are split into text detection and text recognition, sorted by publication date, and each entry carries the reported F1 score on the standard ICDAR benchmarks, so you can compare methods without digging through every paper. The resource columns are where it earns its keep: links to official code, trained models, datasets, and demos sit right next to each paper, which spares you the usual hunt across arXiv and scattered GitHub repos. Since it is a curated index rather than software, think of it as a map for picking a baseline or a recent approach, then follow the linked implementations to actually run something. Newcomers get a clear picture of how scene-text methods evolved, and teams scoping an OCR project can shortlist candidate architectures in one pass.