Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add py impl of pr 2574 #2604

Merged

Conversation

littletomatodonkey
Copy link
Collaborator

score=fast

image

score=slow
image

@MissPenguin
Copy link
Collaborator

good

@MissPenguin MissPenguin merged commit 7b75979 into PaddlePaddle:release/2.1 Apr 23, 2021
@littletomatodonkey littletomatodonkey deleted the 2.1/fix_db_post branch April 25, 2021 12:56
@zcswdt
Copy link

zcswdt commented May 7, 2021

score=fast

试了一下改良以后的效果,对于如下这种弯曲的图片检测,文本框还是有点问题。文本框并不能很好的包络文字区域。
image
我试了别的算法检测效果如下:
image

@buptlihang
Copy link
Contributor

我觉得可能有两个原因导致:

  • 后处理时unclip过度,这可以通过设置unclip_ratio来改善
  • 由于PaddleDB的后处理将所有文本实例都按照矩形文本处理(用了cv.minAreaRect),这对弯曲文本很不友好,弯曲度越高,包络越差,这也可以根据自己的需求修改后处理逻辑。

以及如果从 "检测框最好完整包含整行文本” 这个角度看,我认为PaddleDB的效果其实更好 =) 。

@cw-plus
Copy link

cw-plus commented Apr 15, 2022

score=fast

试了一下改良以后的效果,对于如下这种弯曲的图片检测,文本框还是有点问题。文本框并不能很好的包络文字区域。 image 我试了别的算法检测效果如下: image

请问是哪个检测算法呢

an1018 pushed a commit to an1018/PaddleOCR that referenced this pull request Aug 17, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants