Mingkun Yang

icon indicating a website link https://scholar.google.com/citations?user=3EfF1qgAAAAJ&hl=en icon indicating an email [email protected]

icon company Huazhong University of Science and Technology icon location Luoyu Road 1037, Wuhan, China icon location Hi, I am a Ph.D. student at Huazhong University of Science and Technology. My work focuses on computer vision, especially scene text recognition and text vqa.

Results 23 comments of Mingkun Yang

About chinese text recognition

Yes, with minor modifications, I tried Chinese text recognition based on this repo. You can try it.

How to get character level confidence ??

At present, I don't implement this function with beam search. However, if you don't use the beam search while inference, the character-level recognition score can be obtained via [here](https://github.com/ayumiymk/aster.pytorch/blob/master/lib/utils/labelmaps.py#L45).

how to predict vertical or 90 degree rotated image

In practice, we firstly rotate this image by 90° or -90°. And then select the result with a higher recognition score.