i want to remove any Non Text regions from an captured image as a preprocessing step for OCR Engine, any idea , demo , source code about doing that will be helpful,thanks.
可以将文章内容翻译成中文,广告屏蔽插件可能会导致该功能失效(如失效,请关闭广告屏蔽插件后再试):
问题:
回答1:
I guess this question could be seen as a possible duplicate of your other question: How to detect Text Area from image?
if it wasn't asked in reverse!
Anyway, I rather the other way of thinking about this problem, which is: anything that is not a text region should be ignored.
At this point I need to refer to my other post, which shows an approach to deal with this situation. The idea is that you only send to the OCR Engine the regions that were successfully detected as text, thus ignoring all the non-text regions in the image.