I had problems extracting tabular data from images and PDFs, so I created a Document Structure Detection API - https://extracttable.com
saradhi[@]extracttable