识别率不高，请问下在进行OCR前是否对图像进行处理

apm1467 / videocr

Extract hardcoded subtitles from videos using machine learning

MIT License

506 stars 117 forks source link

Open ducomeon opened 4 years ago

ducomeon commented 4 years ago

作者你好，我想请问下在进行OCR前是否对图像进行处理？还有就是能否增加新的功能，就不并不对每一帧进行识别，设置好时间间隔进行识别？谢谢！

SWHL commented 2 years ago

devmaxxing commented 2 years ago

FYI I've added a frames_to_skip parameter in my fork of this library: https://github.com/oliverfei/videocr-PaddleOCR