SVTR 无法处理长文本条？

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

42.98k stars 7.71k forks source link

SVTR 无法处理长文本条？ #9684

Closed Gavinic closed 1 year ago

Gavinic commented 1 year ago

使用SVTR进行inference的时候，发现模型对于长文本条(文字长度大于50)无法识别，识别效果差。请问有什么解决方案吗？

LDOUBLEV commented 1 year ago

svtr是定长输入，超过25个字符的识别效果会变差，要么你重新训练，加大输入字符长度限制 https://github.com/PaddlePaddle/PaddleOCR/blob/2be9fe763e4ebb55d59cc192eee437b0e09207ce/configs/rec/rec_svtrnet_ch.yml#L18

ccx1997 commented 10 months ago

欢迎关注iccv2023论文 LISTER，解决文本识别字符串长度问题： https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/OCR/LISTER