IssueTracker

We are using Google Cloud Vision (google-cloud-vision 0.37.0) where we have a use case to read and identify characters from pdf/images ( PDFs are created using screenshots ) for review purposes.

The problems we have encountered are listed below:

      - OCR is unable identify superscripts
      -  it is not recognizing symbols ex:  §,¶,±,ǂ, * etc.
      - it is not recognizing bullet points
      - Symbols like dagger (ǂ) is read as 't'
      -  Not extracting text properly when image contain one or more column, text will not be in the same order
     -  if image has hidden text then we cant read the text properly
    -  losing characters
                           if we have 300px resolution images we are missing some chars,
                          same image if we reduce to 150 or 100 we will get chars which are missed in 300pixels resolution but we
                          may loss some other chars
     -  getting additional characters in text block

What you expected to happen:
     - All the above problems needs to be resolved and the images should return the exact characters

Steps to reproduce:
    - Attached image for reproducing the issue


Other information (workarounds you have tried, documentation consulted, etc):
    We couldnt find any work around