I am using form parser currently from DocAI processors. Can you give me more clarity on whether the confidence score shown is from OCR model or any other layout detection that was used in getting the co-ordinates) or the confidence score of the entity-value itself. For generic entities i know we get the confidence properly that this text belongs to that entity with 0.x confidence score. but for other key value pairs what is the scenario?
I have checked the documentation and videos from google cloud community. But it was said that the score mentions how confident the model extracted the text. So I need a clarity, if the confidence is about extracting the text or assigning it to respective entity value. If some number 990998 is present in my doc with certain 0.x as confidence score then i want to know if x is about the confidence of model identified 990998 or is it about the model identified 990998 **as a telephone number **
