search

mjq11302010044 / Real-CE

Real-CE: A Benchmark for Chinese-English Scene Text Image Super-resolution (ICCV2023)
65 stars 6 forks source link

script about extracting annos #5

Open 649459021 opened 7 months ago

649459021 commented 7 months ago

Hi,

Thanks for releasing this amazing TextSR benchmark.

I would like to apply this on other datasets. Could privide the script to extract annos information from images?

Looking forward to your reply~

mjq11302010044 commented 7 months ago

@649459021 There is not a unified script provided. You can simply combine some open-source code text detectors like https://github.com/MhLiao/DB and Chinese OCR tools like https://github.com/FudanVI/benchmarking-chinese-text-recognition.

649459021 commented 7 months ago

Thanks for your reply ~

By the way, which config file is used by the pre-trained model you provided?