Hi Avadhesh,
If you extract text from the PDF with OCR it means you're using region mode (Surface Automation).
Did you already try saving the read area as image, to verify that you are indead reading out the right area that contains the actual data?
If so, did you try increasing the Scale on the read stage which Reads the Text with OCR?
If you're working with scans, verify with previously succesfully handled scans if there are changes in position (Rotations, slight move of) or quality.