Subtitleedit: Missing Slovak language dictionary in Tesseract OCR

[lucybook]

Hi, Slovak language OCR tesseract dictionaries cannot be downloaded via SubtitleEdit and when I try download it from tesseract git it makes SubtitleEdit crash everytime... also their "slk.traineddata" is way bigger than yours.

[Elheym]

Yes, indeed slk.traineddata not available for download from SE. You are not for the 3.02 version downloaded, for this reason to crash everytime... You need this one: https://sourceforge.net/projects/tesseract-ocr-alt/files/tesseract-ocr-3.02.slk.tar.gz/download
Checked - works.

[Elheym]

@niksedk : this many also missed, add please

• Croatian
• Estonian
• Hebrew
• Indonesian
• Latvian
• Malay

[lucybook]

When I say crash i mean this http://imgur.com/txyQgpI

[niksedk]

@lucybook: Could you test latest beta version: https://github.com/SubtitleEdit/subtitleedit/releases/download/3.5.2/SubtitleEditBeta.zip
Better?

[niksedk]

SE still uses Tesseract 3.02 as later versions are not nearly as good... a few things are better, but it mostly just worse and slower.

[darnn]

By the way, does SE train Tesseract, or just use the existing language files?

[niksedk]

SE just uses the existing language files

[Elheym]

@darnn Nope, it does't. Need using existing, but may build yourself. Truth is a troublesome thing...

[lucybook]

@niksedk I tried beta and result is this http://imgur.com/uO5SGjk I have downloaded tesseract slovak dictionary with SE and also slovak spell checking dictionary. All the orange lines are empty btw...

[Elheym]

@lucybook Need to add "eng.traineddata" then it will work.

[niksedk]

New beta up: https://github.com/SubtitleEdit/subtitleedit/releases/download/3.5.2/SubtitleEditBeta.zip
The Slovak 3.02 dictionary seems unfinished, so SE now links to Slovak 3.04 version...

[lucybook]

It's working now, thank you 🥇