Architecture Overview - AI-Powered Health Data Masking

Architecture Overview

Deploying this solution builds the following environment in the AWS Cloud.

        AI-Powered Health Data Masking architecture on AWS

Figure 1: AI-Powered Health Data Masking architecture on AWS

The AWS CloudFormation template deploys an Amazon API Gateway to invoke the microservices (AWS Lambda functions). The microservices provide the business logic to manage preprocessing configuration and logic, and identify and mask health data. The microservices interact with Amazon Rekognition to identify text in an uploaded medical image, and the Amazon Comprehend Medical protected health information data extraction and identification (PHId) API to identify health data in text.

Additionally, the template deploys an Amazon Simple Storage Service (Amazon S3) bucket for storing raw and masked images, AWS CloudTrail to log API actions, and Amazon CloudWatch Logs to log errors within the AWS Lambda functions. By default, log files are encrypted over HTTPS.

For more information on API calls created by the solution, see Appendix A.