5000-Images-Handwriting-OCR-Data-of-Traditional-Chinese-Characters-Taiwan-China

Description

There are 5,000 images of handwriting data of traditional Chinese characters. Texts in the data were annotated for the line-level quadrilateral bounding box. The data can be used for chinese characters recognition application

For more details, please refer to the link: https://www.nexdata.ai/datasets/1190?source=Github

Data size

5,000 images

Collecting environment

including A4 paper, square paper, lined paper, etc.

Device

cellphone

Photographic angle

eye-level angle

Data format

the image data format is .jpg, the annotation file format is .json

Annotation content

line-level quadrilateral bounding box annotation and transcription for the texts

Accuracy

the error bound of each vertex of quadrilateral bounding box is within 5 pixels, which is a qualified annotation, the accuracy of bounding boxes is not less than 97%; the texts transcription accuracy is not less than 97%

Licensing Information

Commercial License