Amazon Rekognition Custom Labels adds support for KMS encryption of custom model training data
Amazon Rekognition Custom Labels supports the ability to use an AWS Key Management Service (KMS) key, to encrypt the image data delivered to the Rekognition Custom Labels service to train your custom machine learning model. By default, Amazon Rekognition Custom Labels encrypts the image data at rest using server-side encryption. With this release, if you provide Amazon Rekognition Custom Labels with your KMS key or alias ARN, Amazon Rekognition Custom Labels will use that KMS key instead to encrypt your training and test images.
Amazon Rekognition Custom Labels supports the ability to use an AWS Key Management Service (KMS) key, to encrypt the image data delivered to the Rekognition Custom Labels service to train your custom machine learning model. By default, Amazon Rekognition Custom Labels encrypts the image data at rest using server-side encryption. With this release, if you provide Amazon Rekognition Custom Labels with your KMS key or alias ARN, Amazon Rekognition Custom Labels will use that KMS key instead to encrypt your training and test images.
To get started, select an existing KMS key or create a new KMS key. When you train a new model, you can then provide the KMS key to Amazon Rekognition Custom Labels by calling the CreateProjectVersion API with your KMS key or alias ARN. The images delivered for training the Amazon Rekognition Custom Labels model will be encrypted at rest using server-side encryption with the given KMS key. If you do not provide Amazon Rekognition Custom Labels with a KMS key or alias ARN, then Amazon Rekognition Custom Labels will use the default KMS key to encrypt the data.
Support for this new feature is automatically enabled and available in all Amazon Rekognition Custom Labels regions. To learn more about Amazon Rekognition Custom Labels and region availability, please visit our documentation and region table.
Get started with Amazon Rekognition Custom Labels today.