Character coordinates in Textract

Question

Hello Everybody,  
  
I'm working with textract to OCR some documents, I am able to access only words and sentences coordinates and bounding boxes. Is there anyway I can get the coordinates of the each character of the word?

Answer

Hi,   
Thanks for using AWS Textract. Currently we don't support providing character coordinates. We will take your feedback and evaluate the feasibility.  
  
thanks,  
Fang

Answer

Hello Fang,  
  
Thanks for the reply. I was actually looking for character coordinates to save the layout data of entire document to the Character level  because if I have character level information I can do fuzzy extraction and swap the characters accordingly if there are any errors.  
  
Edited by: geenome on Jan 20, 2021 2:34 PM

Answer

Hi,  
Thanks for providing your detailed use case. This is helpful for our service.  
Fang

Character coordinates in Textract

Contenido relevante