Does Document AI not detect the shopping receipt?

Question

I can successfully extract the details of the shopping receipt (products, store name, etc.) with the expense parser in Document AI. I wrote the store name and product name in handwriting on normal paper and uploaded it, but it still took out the information. Does Document AI not understand if this is a shopping receipt?

Answer 1

Yes, Document AI recognizes that it is a shopping receipt. Document AI just extracts the data of the receipt that is not handwritten. If you want to upload a handwritten receipt, you need to upload an image, PDF or TIFF files, which are the formats in which the OCR recognizes handwriting text.

Document AI is a document understanding solution that takes unstructured data (documents, forms, etc.) and makes the data easier to understand, analyze, and consume by providing structure through content classification, entity extraction, advanced searching, and more.

An option could be to use AutoML Vision Object Detection to convert sections of images into text documents.

Answer 2

Document AI OCR can support handwriting in any supported file format.

The Expense Parser is a specialized processor trained for parsing receipts, however it may not perform as expected for non-printed receipts (since most receipts are printed in similar formats).

I recommend reviewing over the documentation for handling the processing response for specialized processors.

You can also try out the Specialized Processors codelab which demonstrates how to extract entities for any specialized processor.

The documentation also contains a Sample Receipt PDF and full output from the Expense Parser.