提交历史

作者 SHA1 备注 提交日期
  Michele Yin cc793b3616 initial ocr support for mps devices 10 月之前
  thatDudo d8009c5800 Experimental: Use ctd mask as raw_mask for default mask refiner 1 年之前
  thatDudo 723c43d5ec Move textline merge after ocr 1 年之前
  thatDudo 6a9e4b0362 Add "none" detector (Closes #364) 1 年之前
  zyddnys a0640c5166 bugfix 1 年之前
  thatDudo 45d2f379ec Add --det-invert and --det-rotate which will apply filter for detection 1 年之前
  thatDudo aeb32d4b71 Rename ModelWrapper.forward to infer to avoid confusion with pytorch 1 年之前
  thatDudo d1190ae59f Make ctd use its own mask_refinement due to better results 1 年之前
  thatDudo aa271762c2 Remove debug prints 1 年之前
  thatDudo 0437e89c82 Improve text_region sorting 1 年之前
  thatDudo 942515192c Add simple text_region sorting after detection. 1 年之前
  thatDudo 374ec0ed66 Add --model-dir argument 1 年之前
  thatDudo 0cddb85039 [renderer] Introduce minimum bbox size 1 年之前
  thatDudo bd48e35e92 Refactor TextBlock class 1 年之前
  thatDudo 65571e50d2 Make ctd detector use common textline merge and mask refinement 1 年之前
  thatDudo 548e6ef657 Turn utils.py into submodule and move textblock.py into there 1 年之前
  thatDudo f4b7994b4b Refactor translate_demo and restructure project 1 年之前