I'm using TessBaseAPI to scan a png of WIndows screenshot,
Issue 1) List of numbers are inaccurate
7254134516423432 1324152643132424 324326178176892
text out=> 72541345 16423432 1324152643132424 324326178176892
Tesseract put a space between 5 and 1
2) Missing numbers
If there are 2 notepad windows in screenshot it fails to get the 2nd list, just fragments
Input File is a png high quality. I increase resolution to 300%.
I tried setting textord_space_size_is_variable to 1
and
I tried setting tosp_min_sane_kn_sp to different values but no difference,
How can I get the best results?
Thanks