Architecture Overview
Deploying this solution builds the following environment in the AWS Cloud.
Figure 1: Auto Check-In App architecture
The AWS CloudFormation
When an attendee registers for your event, they can upload a photo which is stored
in an Amazon S3 bucket. The upload to Amazon S3 triggers a Lambda function that calls
the Amazon Rekognition IndexFaces API. Amazon Rekognition extracts facial features into a feature vector and creates a
face_id. Then, the vector is stored in a face collection, and the face_id and corresponding user name are stored in a DynamoDB table.
At the event, an operator uses the solution’s Python-based UI and a camera to take
a picture of the attendee, crop the photo, and send the photo to Amazon API Gateway
which triggers a Lambda function that calls the Amazon Rekognition SearchFacesByImage API. Amazon Rekognition extracts the facial features from the image into a feature vector
and compares the vector to the vectors in the face collection. When Amazon Rekognition
finds a face with high similarity, the face_id is used to retrieve the user name. The user name is sent to the operator’s laptop
showing that the attendee has been authenticated.
