How do I remove extra unnecessary characters from OCR output

[karra...@gmail.com]

Hi,

I'm using OCR for my Raspberry Pi book reader project and I seem to have this issue where the OCR produces extra characters along with the text every time. I tried using some of the config parameters to see if it makes any difference which hasn't been the case so far. 

For example, here is my output for the following text from a Dr.Seuss book: (The text that was on the page was just Everyone wants

a big green kangaroo.) How can I get rid of the extra characters? Thanks

4441‘ 'muw-u»

A; Wmme

\ K r'f'.

,\

3 I ,

E v \y r

3 \g\\ A  '

'3 I» ’1,” ‘

 ,x

i, 1/ /

I

¢ 1

“.

/ ‘x

m- WWW-““a”..-

  Everyone wants

a big green kangaroo.

: is C-Bpfright Cnnwnucns.

r-wwrwmlaai.vn~is&‘-M'ikiww. _ . A

[TAPAN SHARMA]

It may be possible due to compression artifacts in Image. You can achieve a lot of improvement using lossless compression like PNG than JPEG.