02 April 2018
Unfortunately, this is a known issue: our current OCR technology does not recognize such narrow tall text as here properly. This is a systematic issue, and the extraction of such text should be improved in the next version of ABBYY OCRT. However we should not expect this fix very soon, because transferring to the new version of OCRT is planned, but not in the nearest future.
Do you have possibility to resize the image vertically thereby artificially making the text not so narrow and tall? When I have resized your image in percentage terms 100 x 70 and processed it in Cloud OCR SDK using the following URL:
the output is quite accurate:
Kindly try this, hope these tips will be useful!