Not able to extract Certain field from a document

Question

I want to extract gross income from Income-tax (a document).Here are the things i have tried. 1)Using Ctpn ,form the countours,and crop those parts,and fed it to pytesseract.But it is giving very bad results. 2) I tried Adaptive thersolding , before feeding to to pytesseract, doesn't helped. 3)The image should be greater than 70 dpi before we are feeding to tesseract , so I am scalling it too . Additionally i have tried taking average but hard luck.

Can anyone please tell me how can i extract that.