overlapping boxes displayed for Devanagari script

[Shreeshrii]

Please see attached - there are multiple boxes displayed for devanagari script even though there is one entry in box file ...

[Shreeshrii]

Sample Box Tiff pairs for Sanskrit - san language, Devanagari script can be found at
https://github.com/Shreeshrii/imagessan/tree/master/san95-box-tiff

Corresponding traineddata is at https://github.com/Shreeshrii/imagessan/blob/master/tessdata/san95.traineddata

You could use it for testing for tesseract4java performance with complex scripts.

Thanks!

[Shreeshrii]

I am wondering whether the overlap is because of multi-page tifs and corresponding box files.

[Shreeshrii]

https://sourceforge.net/projects/vietocr/files/jTessBoxEditor/ has implementation of editor for multipage tiffs.