OCR Datasets

15,148 Sets - Japanese and Korean Menu OCR Data, including 9,701 Japanese menus and 5,447 Korean menus, annotation content includes row level quadrilateral box annotation, row level content transcription (a small portion of data is column level quadrilateral box annotation, column level content transcription)

This dataset contains 1,586,458 document data sets covering a wide range of Chinese-language materials, including Chinese textbooks, e-books, and teaching reference materials. The annotation files provide both OCR annotations and structured document analysis. The dataset is suitable for document OCR, document understanding tasks.

102,870 Sets - Chinese Instruction Manual OCR&Parsing Data, including Drug instructions, Installation instructions, Operating Instructions, Product manual, User manual, the annotation files contain OCR annotations and structured parsing.

This dataset includes 9,574 handwriting images across 8 languages, including English, Spanish Portuguese and more. The data diversity includes multiple collecting scenes, different text carriers and different photographic angles(looking up, eye-level, looking down). In terms of annotation, each text line is annotated with quadrilateral polygons and transcription. The dataset can be used for training and evaluating OCR models, handwriting recognition systems, and multilingual text extraction tasks in AI and computer vision.

5,000 Images-4 Categories Cards OCR Data, applicable for tasks such as character recognition.

This dataset contains 128,900 images featuring text in seven languages: Arabic, French, German, Hindi, Italian, Portuguese, and Spanish. The data covers diverse real-world scene conditions, including blurred and natural scenes, as well as challenging capture conditions such as handheld shooting and moiré patterns. Text regions are annotated with polygonal bounding boxes for precise text localization, together with corresponding text transcriptions. This dataset can be used for multilingual OCR, scene text detection, and OCR model training.

This dataset contains 39,993 images, primarily featuring Chinese text with a small portion of English text. The dataset includes line-level rectangular bounding box annotations and corresponding text transcriptions. A small portion of the data additionally provides column-level quadrilateral bounding boxes and text transcriptions for more precise localization of text regions. The dataset can be used for OCR, text detection, text recognition, OCR model training, and document image understanding tasks.

This dataset contains 4,995 Vietnamese OCR images with annotations and text transcriptions. The data includes 258 natural scene images, 2,553 Internet images, and 2,184 document images. For line-level content annotation, quadrilateral bounding box annotations and text transcriptions are provided. For column-level content annotation, column-level quadrilateral bounding box annotation and text transcription are provided. The data can be used for tasks such as Vietnamese recognition in multiple scenes.

14,980 Images PPT OCR Data of 8 Languages. This dataset includes 8 languages, multiple scenes, different photographic angles, different photographic distances, different light conditions. For annotation, line-level quadrilateral bounding box annotation and transcription for the texts were annotated in the data. The dataset can be used for tasks such as OCR of multi-language.