Custom Dataset adequate annotation for text detection and recognition

PaddlePaddle / PaddleOCR

Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and deployment among server, mobile, embedded and IoT devices)

https://paddlepaddle.github.io/PaddleOCR/

Apache License 2.0

43.81k stars 7.79k forks source link

Custom Dataset adequate annotation for text detection and recognition #9593

Closed mahachaaben99 closed 5 months ago

mahachaaben99 commented 1 year ago

Hello gyus , I have a custom dataset and I woud like to know exactly the annotation format I should follow while labeling my dataset in order to train PaddleOCR models on it

sumitclearquote commented 1 year ago

@maziqueen79 Could you tell me what the appropriate annotation format is here?

UserWangZz commented 5 months ago

This issue has not been updated for a long time. This issue is temporarily closed and can be reopened if necessary.