Transkribus / TranskribusCore

Note: the repo has been moved to https://gitlab.com/readcoop/Transkribus/TranskribusCore
GNU General Public License v3.0
37 stars 5 forks source link

Text2Image tool returns empty transcription #51

Open TommasoBendinelli opened 4 years ago

TommasoBendinelli commented 4 years ago

Hi there! I am playing around with your tool and I tried to automatically transcribe a document of mine. Then I exported the document with the txt option and imported again to test out the Text2Image tool.

When importing the txt and images together I get lines in the trascription widget but of course they are not linked in the canvas. I then try to run the Text2Image tool with the default settings.

Unfortunately the output of the tool is an empty transcription. Is it me doing something wrong or there is an underlying issue? I tried with both English Writing M1 as base model and English Writing M2.

I also followed the same procedure with the file "English_Handwriting 0.1", where I get good results.

Best, Tommaso

TommasoBendinelli commented 4 years ago

Might be that the input picture format is playing a part? (Although I find it not probable as the segmentation is being performed correctly on my document )