I have this simple image with a date:
Tesseract produces the output:
$ tesseract test.png -
Estimating resolution as 233
03:41 pm
In similar images, I have the problem that it misunderstands 1's for 7's and the other way around. How can I help Tesseract to recognise these characters?
My version of Tesseract is:
$ tesseract -v
tesseract 5.0.0-alpha-20210401-130-g7a308
leptonica-1.79.0
libgif 5.1.4 : libjpeg 8d (libjpeg-turbo 2.0.3) : libpng 1.6.37 : libtiff 4.1.0 : zlib 1.2.11 : libwebp 0.6.1 : libopenjp2 2.3.1
Found AVX2
Found AVX
Found FMA
Found SSE4.1
Found OpenMP 201511