For AWS Custom Face Rekognition to be installed, it is a requirement that your Curator system is of version 3.3 or later. If you are using any version prior to 3.3, this installation will not work as expected. In addition to this, ensure you have Device Director version 5.17.1.5 or beyond installed and check that you have Process Engine Service 2.16.1.16 or beyond installed.
You will need to install the AWS Rekognition module version 1.2.1 via the Process Engine Modules tab.
Next, create/modify a send to destination plugin, so that you have two options to select:
- An option named Scan File for Custom Faces, for which the destination mediastore is AWS-REKOGNITION-FACES
- A second option named Scan Collection to Update Facial Rekognition Library, for which the destination store is AWS-REKOGNITION-UPDATE-FACE-COLLECTION
Next, some metadata must be manually created in Curator System Administrator.
Head to CSA Metadata Names and add:
Name: FaceCollectionLabel
Display Name: FaceCollectionLabel
Asset type: Media
Data Type: Text
Grant all workgroups to the metadata which will be enabled for this workflow.
Once you have completed the above instructions, the installation should be complete. Please refer to the AWS Custom Face Rekognition User Guide to understand how to test and use the plugins/workflow.