Summary
In this chapter, we discussed the core features of Amazon Comprehend, PII detection and redaction, and Amazon Comprehend Medical’s PHI detection feature. We also discussed the Review and Validation stage of the IDP pipeline and why it is critical for accurate IDP. We also discussed how to leverage Amazon Textract to extract text from any type of document and then pass it to Amazon Comprehend (Medical) for PII or PHI information detection and redaction. This helps to build a document processing pipeline to handle sensitive information.
We then reviewed the need for human review. We also discussed Amazon A2I and its core features for including human beings in the review of more critical field elements in documents, or ones with lower accuracy. This automation helps build cost-effective document processing with time acceleration.
In the next chapter, we will discuss how to build a data lake for health information and how IDP can be integrated with Amazon HealthLake...
