How does Tesseract OCR fuse results from multiple models?

tesseract-ocr / tessdata_best

Best (most accurate) trained LSTM models.

Apache License 2.0

1.26k stars 387 forks source link

How does Tesseract OCR fuse results from multiple models? #81

Open yaofuzhou opened 8 months ago

yaofuzhou commented 8 months ago

In the latest Tesseract, the -l option allows for, say, eng+chi_sim. How does Tesseract OCR fuse results from multiple models? Can someone point to the code that does the fusion? Thanks!

inahwangg commented 4 months ago

저도 궁금했는뎅, please check this :)

doc/combine_tessdata.1.asc
doc/combine_lang_model.1.asc
doc/lstmtraining.1.asc